Entity

Time filter

Source Type

Pittsburgh, PA, United States

Carnegie Mellon University is a private research university in Pittsburgh, Pennsylvania.The university began as the Carnegie Technical Schools founded by Andrew Carnegie in 1900. In 1912, the school became the Carnegie Institute of Technology and began granting four-year degrees. In 1967, the Carnegie Institute of Technology merged with the Mellon Institute of Industrial Research to form Carnegie Mellon University. The university's 140-acre main campus is 3 miles from Downtown Pittsburgh and abuts the Carnegie Museums of Pittsburgh, the main branch of the Carnegie Library of Pittsburgh, the Carnegie Music Hall, Schenley Park, Phipps Conservatory and Botanical Gardens, the Pittsburgh Golf Club, and the campus of the University of Pittsburgh in the city's Oakland and Squirrel Hill neighborhoods, partially extending into Shadyside.Carnegie Mellon has seven colleges and independent schools: the Carnegie Institute of Technology , College of Fine Arts, Dietrich College of Humanities and Social science, Mellon College of Science, Tepper School of Business, H. John Heinz III College and the School of Computer Science. Carnegie Mellon fields 17 varsity athletic teams as part of the University Athletic Association conference of the NCAA Division III. Wikipedia.


Acharya A.,Carnegie Mellon University
Journal of Elasticity | Year: 2011

A methodology is devised to utilize the statistical mechanical entropy of an isolated, constrained atomistic system to define constitutive response functions for the dissipative driving-force and energetic fields in continuum thermomechanics. A thermodynamic model of dislocation mechanics is discussed as an example. Primary outcomes are constitutive relations for the back-stress tensor and the Cauchy stress tensor in terms of the elastic distortion, mass density, polar dislocation density, and the scalar statistical density. © Springer Science+Business Media B.V. 2011.


Christou I.T.,Athens Information Technology | Christou I.T.,Carnegie Mellon University
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2011

We present a novel optimization-based method for the combination of cluster ensembles for the class of problems with intracluster criteria, such as Minimum-Sum-of-Squares-Clustering (MSSC). We propose a simple and efficient algorithmcalled EXAMCEfor this class of problems that is inspired from a Set-Partitioning formulation of the original clustering problem. We prove some theoretical properties of the solutions produced by our algorithm, and in particular that, under general assumptions, though the algorithm recombines solution fragments so as to find the solution of a Set-Covering relaxation of the original formulation, it is guaranteed to find better solutions than the ones in the ensemble. For the MSSC problem in particular, a prototype implementation of our algorithm found a new better solution than the previously best known for 21 of the test instances of the 40-instance TSPLIB benchmark data sets used in [CHECK END OF SENTENCE], [CHECK END OF SENTENCE], and [CHECK END OF SENTENCE], and found a worse-quality solution than the best known only five times. For other published benchmark data sets where the optimal MSSC solution is known, we match them. The algorithm is particularly effective when the number of clusters is large, in which case it is able to escape the local minima found by K-means type algorithms by recombining the solutions in a Set-Covering context. We also establish the stability of the algorithm with extensive computational experiments, by showing that multiple runs of EXAMCE for the same clustering problem instance produce high-quality solutions whose Adjusted Rand Index is consistently above 0.95. Finally, in experiments utilizing external criteria to compute the validity of clustering, EXAMCE is capable of producing high-quality results that are comparable in quality to those of the best known clustering algorithms. © 2011 IEEE.


Settles B.,Carnegie Mellon University
EMNLP 2011 - Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference | Year: 2011

This paper describes DUALIST, an active learning annotation paradigm which solicits and learns from labels on both features (e.g., words) and instances (e.g., documents). We present a novel semi-supervised training algorithm developed for this setting, which is (1) fast enough to support real-time interactive speeds, and (2) at least as accurate as preexisting methods for learning with mixed feature and instance labels. Human annotators in user studies were able to produce near-state-of-the-art classifiers on several corpora in a variety of application domains with only a few minutes of effort. © 2011 Association for Computational Linguistics.


Natarajan A.,Carnegie Mellon University
Advances in Astronomy | Year: 2011

We discuss the formation of dark matter caustics, and their possible detection by future dark matter experiments. The annual modulation expected in the recoil rate measured by a dark matter detector is discussed. We consider the example of dark matter particles with a Maxwell-Boltzmann velocity distribution modified by a cold stream due to a nearby caustic. It is shown that the effect of the caustic flow is potentially detectable, even when the density enhancement due to the caustic is small. This makes the annual modulation effect an excellent probe of inner caustics. We also show that the phase of the annual modulation at low recoil energies does not constrain the particle mass unless the velocity distribution of particles in the solar neighborhood is known. © 2011 Aravind Natarajan.


Khair A.S.,Carnegie Mellon University
Journal of Rheology | Year: 2016

An asymptotic solution to the Giesekus constitutive model of polymeric fluids under homogenous, oscillatory simple shear flow at large Weissenberg number, Wi 蠑 1, and large strain amplitude, Wi/De 蠑 1, is constructed. Here, Wi = λ∗γ∗0, where λ∗ is the polymer relaxation time and γ∗0 is the shear rate amplitude, and De = λ∗ω∗ is a Deborah number, where ω∗ is the oscillation frequency. Under these conditions, we show that the first normal stress difference is the dominant rheological signal, scaling as G∗Wi1/2, where G∗ is the elastic modulus. The shear stress and second normal stress difference are of order G∗. The polymer stress exhibits pronounced nonlinear oscillations, which are partitioned into two temporal regions: (i) A "core region," on the time scale of λ∗, reflecting a balance between anisotropic drag and orientation of polymers in the strong flow; and (ii) a "turning region," centered around times when the shear flow reverses, whose duration is on the hybrid time scale (λ∗)2/3/(γ∗0)1/3, in which flow-driven orientation, anisotropic drag, and relaxation are all leading order effects. Our asymptotic solution is verified against numerical computations, and a qualitative comparison with relevant experimental observations is presented. Our results can, in principle, be employed to extract the nonlinearity (anisotropic drag) parameter, α, of the Giesekus model from experimental data, without the need to fit the stress waveform over a complete oscillation cycle. Finally, we discuss our findings in relation to recent work on shear banding in oscillatory flows. © 2016 The Society of Rheology.


Williams W.O.,Carnegie Mellon University
Journal of Elasticity | Year: 2011

Huxley's cross-bridge dynamics of muscle contraction is widely used in understanding, in particular, laboratory experiments on muscles and subunits of muscle. The hardconnection version of the model has several defects. In this paper I present a detailed and precise method of solution of the problem with a compliant element in series with the muscle. © The Author(s) 2011.


Nagle J.F.,Carnegie Mellon University
Faraday Discussions | Year: 2012

One of the many aspects of membrane biophysics dealt with in this Faraday Discussion regards the material moduli that describe energies at a supramolecular level. This introductory lecture first critically reviews differences in reported numerical values of the bending modulus KC, which is a central property for the biologically important flexibility of membranes. It is speculated that there may be a reason that the shape analysis method tends to give larger values of KC than the micromechanical manipulation method or the more recent X-ray method that agree very well with each other. Another theme of membrane biophysics is the use of simulations to provide exquisite detail of structures and processes. This lecture critically reviews the application of atomic level simulations to the quantitative structure of simple single component lipid bilayers and diagnostics are introduced to evaluate simulations. Another theme of this Faraday Discussion was lateral heterogeneity in biomembranes with many different lipids. Coarse grained simulations and analytical theories promise to synergistically enhance experimental studies when their interaction parameters are tuned to agree with experimental data, such as the slopes of experimental tie lines in ternary phase diagrams. Finally, attention is called to contributions that add relevant biological molecules to bilayers and to contributions that study the exciting shape changes and different non-bilayer structures with different lipids. This journal is © The Royal Society of Chemistry.


Tomich P.L.,Kent State University | Helgeson V.S.,Carnegie Mellon University
Journal of Traumatic Stress | Year: 2012

This study examined the linkage of posttraumatic growth (PTG) to quality of life (QOL) among individuals newly diagnosed with cancer. Individuals (26 men, 36 women) reported PTG 3 months postdiagnosis (T1) and 3 months later (T2). Cross-sectional analyses revealed a linear association between PTG and QOL-more PTG was related to worse mental health at T1 (β = -28). PTG, however, revealed a quadratic relationship with depressive symptoms at T1 and physical health at T2: Individuals with high or low levels of PTG had fewer depressive symptoms and better QOL than those with moderate levels. Longitudinal analyses revealed a linear association between PTG and QOL; more PTG at T1 predicted better physical health at T2. There were no longitudinal curvilinear associations. Although the linear links of PTG to QOL were contradictory within this study, both of the curvilinear relations, although not robust, confirm previous research. Further analyses differentiated low, medium, and high PTG groups in terms of perceiving cancer as stressful, intrusive thoughts, and coping strategies. Overall, relations of PTG to adjustment may be more complex and dynamic than previously assumed. Clinicians should consider the notion that more growth may sometimes, but not always, be better. © 2012 International Society for Traumatic Stress Studies.


Rajkumar R.,Carnegie Mellon University
Proceedings of the IEEE | Year: 2012

Cyber-physical systems (CPSs) couple the cyber aspects of computing and communications tightly with the dynamics and physics of physical systems operating in the world around us. This emerging multidisciplinary frontier will enable revolutionary changes in the way humans live. Just like the Internet has transformed how national economies are intertwined, how humans interact, and how commerce is conducted, CPSs will transform how humans interact with and control the physical environment to the greater benefit of society. Major economic sectors that will see dramatic advances will include transportation, energy, buildings, healthcare, manufacturing, physical infrastructure, agriculture, and defense among others. The CPS community will comprise computer scientists, engineers of all stripes as well as many biologists, chemists, and physicians. The arts will also leverage the innovations of CPSs and inspire even more novel inventions in the scientific and engineering realms that combine the world of computing with the physical world. In short, the cyber-physical world of the future will be both very different and very welcome. © 2012 IEEE.


Ettensohn C.A.,Carnegie Mellon University
Evolution and Development | Year: 2014

It is widely accepted that biomineralized structures appeared independently in many metazoan clades during the Cambrian. How this occurred, and whether it involved the parallel co-option of a common set of biochemical and developmental pathways (i.e., a shared biomineralization "toolkit"), are questions that remain unanswered. Here, I provide evidence that horizontal gene transfer supported the evolution of biomineralization in some metazoans. I show that Msp130 proteins, first described as proteins expressed selectively by the biomineral-forming primary mesenchyme cells of the sea urchin embryo, have a much wider taxonomic distribution than was previously appreciated. Msp130 proteins are present in several invertebrate deuterostomes and in one protostome clade (molluscs). Surprisingly, closely related proteins are also present in many bacteria and several algae, and I propose that msp130 genes were introduced into metazoan lineages via multiple, independent horizontal gene transfer events. Phylogenetic analysis shows that the introduction of an ancestral msp130 gene occurred in the sea urchin lineage more than 250 million years ago and that msp130 genes underwent independent, parallel duplications in each of the metazoan phyla in which these genes are found. © 2014 Wiley Periodicals, Inc.


Yagan O.,Carnegie Mellon University
IEEE Transactions on Information Theory | Year: 2016

We introduce a new random key predistribution scheme for securing heterogeneous wireless sensor networks. Each of the n sensors in the network is classified into $r$ classes according to some probability distribution μ=μ1,⋯,μr. Before deployment, a class- i sensor is assigned Ki cryptographic keys that are selected uniformly at random from a common pool of P keys. Once deployed, a pair of sensors can communicate securely if and only if they have a key in common. We model the communication topology of this network by a newly defined inhomogeneous random key graph. We establish scaling conditions on the parameters P and K1,⋯, Kr so that this graph: 1) has no isolated nodes and 2) is connected, both with high probability. The results are given in the form of zero-one laws with the number of sensors n growing unboundedly large; critical scalings are identified and shown to coincide for both graph properties. Our results are shown to complement and improve those given by Godehardt et al. and Zhao et al. for the same model, therein referred to as the general random intersection graph. © 1963-2012 IEEE.


Hackney D.D.,Carnegie Mellon University
Methods in Enzymology | Year: 2012

It is often possible to obtain a detailed understanding of the forward steps in ATP hydrolysis because they are thermodynamically favored and usually occur rapidly. However, it is difficult to obtain the reverse rates for ATP resynthesis because they are thermodynamically disfavored and little of their product, ATP, accumulates. Isotopic exchange reactions provide access to these reverse reactions because isotopic changes accumulate over time due to multiple reversals of hydrolysis, even in the absence of net resynthesis of significant amounts of ATP. Knowledge of both the forward and reverse rates allows calculation of the free energy changes at each step and how it changes when coupled to an energy-requiring conformational step such as unwinding of an RNA helix. This chapter describes the principal types of oxygen isotopic exchange reactions that are applicable to ATPases, in general, and helicases, in particular, their application and their interpretation. © 2012 Elsevier Inc. All rights reserved.


Lang D.,Carnegie Mellon University
Astronomical Journal | Year: 2014

The Wide-field Infrared Survey Explorer (WISE) satellite observed the full sky in four mid-infrared bands in the 2.8-28 μm range. The primary mission was completed in 2010. The WISE team has done a superb job of producing a series of high-quality, well-documented, complete data releases in a timely manner. However, the "Atlas Image" coadds that are part of the recent AllWISE and previous data releases were intentionally blurred. Convolving the images by the point-spread function while coadding results in "matched-filtered" images that are close to optimal for detecting isolated point sources. But these matched-filtered images are sub-optimal or inappropriate for other purposes. For example, we are photometering the WISE images at the locations of sources detected in the Sloan Digital Sky Survey through forward modeling, and this blurring decreases the available signal-to-noise by effectively broadening the point-spread function. This paper presents a new set of coadds of the WISE images that have not been blurred. These images retain the intrinsic resolution of the data and are appropriate for photometry preserving the available signal-to-noise. Users should be cautioned, however, that the W3- and W4-band coadds contain artifacts around large, bright structures (large galaxies, dusty nebulae, etc.); eliminating these artifacts is the subject of ongoing work. These new coadds, and the code used to produce them, are publicly available at http://unwise.me. © 2014. The American Astronomical Society. All rights reserved.


Elia N.,Iowa State University | Eisenbeis J.N.,Carnegie Mellon University
IEEE Transactions on Automatic Control | Year: 2011

In this paper, we investigate control across stochastic dropout channels. In particular, we consider the Mean Square Stability of a SISO plant in the case there is only one channel in the feedback loop and the case where both actuator and sensor channels are present. We seek optimal networked control schemes that are memoryless functions of channel state information and for each channel state are otherwise linear and time invariant functions of channel output. We establish a fundamental limit on the dropout probability allowable for the Mean Square Stability of the closed loop system. The maximal tolerable dropout probability is only a function of the unstable eigenvalues of the plant. When either the actuator or the sensor channel is present, we propose a receiver structure that can stabilize the system under the worst dropout probability; moreover, we can simultaneously design the optimal controller and receiver and show that they can be implemented in physically separated locations (decentralized). When both actuator and sensor channels are present in the loop, the main result is a centralized stabilization technique that always achieves the fundamental bound via noiseless acknowledgement from the actuation receiver. Finally, we extend the results to the more general case where also the acknowledgements are lost with a given probability and compute how the unreliable delivery of the acknowledgements affects the minimal quality of service required of the actuator and sensor channels. © 2006 IEEE.


Shimada K.,Carnegie Mellon University
Journal of Computing and Information Science in Engineering | Year: 2011

This paper presents the current issues and trends in meshing and geometric processing, core tasks in the preparation stage of computational engineering analyses. In product development, computational simulation of product functionality and manufacturing process have contributed significantly toward improving the quality of a product, shortening the time-to-market and reducing the cost of the product and manufacturing process. The computational simulation can predict various physical behaviors of a target object or system, including its structural, thermal, fluid, dynamic, and electro-magnetic behaviors. In industry, the computer-aided engineering (CAE) software packages have been the driving force behind the ever-increasing usage of computational engineering analyses. While these tools have been improved continuously since their inception in the early 1960s, the demand for more complex computational simulation has grown significantly in recent years, creating some major shortfalls in the capability of current CAE tools. This paper first discusses the current trends of computational engineering analyses and then focuses on two areas of such shortfalls: meshing and geometric processing, critical tasks required in the preparation stage of engineering analyses that use common numerical methods such as the finite element method and the boundary element method. © 2011 American Society of Mechanical Engineers.


Minshew N.J.,University of Pittsburgh | Keller T.A.,Carnegie Mellon University
Current Opinion in Neurology | Year: 2010

Purpose of Review: Functional magnetic resonance imaging studies have had a profound impact on the delineation of the neurobiologic basis for autism. Advances in fMRI technology for investigating functional connectivity, resting state connectivity, and a default mode network have provided further detail about disturbances in brain organization and brain-behavior relationships in autism to be reviewed in this article. Recent Findings: Recent fMRI studies have provided evidence of enhanced activation and connectivity of posterior, or parietal-occipital, networks and enhanced reliance on visuospatial abilities for visual and verbal reasoning in high functioning individuals with autism. Evidence also indicates altered activation in frontostriatal networks for cognitive control, particularly involving anterior cingulate cortex, and altered connectivity in the resting state and the default mode network. The findings suggest that the specialization of many cortical networks of the human brain has failed to develop fully in high functioning individuals with autism. Summary: This research provides a growing specification of to the neurobiologic basis for this complex syndrome and for the co-occurrence of the signs and symptoms as a syndrome. With this knowledge has come new neurobiologically based opportunities for intervention. © 2010 Wolters Kluwer Health | Lippincott Williams & Wilkins.


Sou S.-I.,National Cheng Kung University | Tonguz O.K.,Carnegie Mellon University
IEEE Transactions on Vehicular Technology | Year: 2011

In vehicular ad hoc network (VANET) safety applications, the source vehicle that detects an accident can generate a warning message and propagate it to the following vehicles to notify other drivers before they reach the potential danger zone on the road. Recent studies have shown that sparse vehicle traffic leads to network fragmentation, which poses a crucial research challenge for safety applications. In this paper, we analyze and quantify the improvement in VANET connectivity when a limited number of roadside units (RSUs) are deployed and to investigate the routing performance for broadcast-based safety applications in this enhanced VANET environment. Our results show that, even with a small number of RSUs, the performance in terms of the probability of network connectivity, the rehealing delay, the number of rehealing hops, and the message penetration time can be significantly improved in highway VANET scenarios. © 2011 IEEE.


Lerner J.S.,Harvard University | Li Y.,University of California at Riverside | Valdesolo P.,Claremont McKenna College | Kassam K.S.,Carnegie Mellon University
Annual Review of Psychology | Year: 2015

A revolution in the science of emotion has emerged in recent decades, with the potential to create a paradigm shift in decision theories. The research reveals that emotions constitute potent, pervasive, predictable, sometimes harmful and sometimes beneficial drivers of decision making. Across different domains, important regularities appear in the mechanisms through which emotions influence judgments and choices. We organize and analyze what has been learned from the past 35 years of work on emotion and decision making. In so doing, we propose the emotion-imbued choice model, which accounts for inputs from traditional rational choice theory and from newer emotion research, synthesizing scientific models. © 2015 by Annual Reviews. All rights reserved.


Blum A.,Carnegie Mellon University | Ligett K.,California Institute of Technology | Roth A.,University of Pennsylvania
Journal of the ACM | Year: 2013

In this article, we demonstrate that, ignoring computational constraints, it is possible to release synthetic databases that are useful for accurately answering large classes of queries while preserving differential privacy. Specifically, we give a mechanism that privately releases synthetic data useful for answering a class of queries over a discrete domain with error that grows as a function of the size of the smallest net approximately representing the answers to that class of queries. We show that this in particular implies a mechanism for counting queries that gives error guarantees that grow only with the VC-dimension of the class of queries, which itself grows at most logarithmically with the size of the query class. We also show that it is not possible to release even simple classes of queries (such as intervals and their generalizations) over continuous domains with worst-case utility guarantees while preserving differential privacy. In response to this, we consider a relaxation of the utility guarantee and give a privacy preserving polynomial time algorithm that for any halfspace query will provide an answer that is accurate for some small perturbation of the query. This algorithm does not release synthetic data, but instead another data structure capable of representing an answer for each query. We also give an efficient algorithm for releasing synthetic data for the class of interval queries and axis-aligned rectangles of constant dimension over discrete domains. © 2013 ACM.


LoBue V.,Rutgers University | Rakison D.H.,Carnegie Mellon University
Developmental Review | Year: 2013

Fear is one of our most basic emotions. It is an important social signal and alerts us to when a situation is safe or risky. Interestingly, not all fears are created equal: Several researchers have proposed that humans develop specific fears, such as fear of threatening stimuli, more readily than others. Here we discuss three major theories of fear acquisition, and consider the possibility that some fears are privileged in learning. Second, we review a growing literature that suggests that humans have perceptual biases that quickly draw attention to threatening stimuli in the environment. In particular, we highlight recent developmental work that shows that even infants and young children respond rapidly to the presence of threat well before they acquire any threat-relevant fears. Finally, we argue that such biases may play a causal role in privileging fear learning for certain threats, and we suggest directions for future work that can clarify whether early biases in perception indeed facilitate the development of our most common fears. © 2013 Elsevier Inc.


Atsumi T.,Hitachi Ltd. | Messner W.C.,Carnegie Mellon University
IEEE Transactions on Industrial Electronics | Year: 2012

We have developed a user-friendly loop-shaping method employing the Robust Bode (RBode) plot for optimizing the head-positioning system in a hard disk drive (HDD). The RBode plot represents the robust performance criterion as allowable and forbidden regions on the open-loop Bode plot. Using the RBode plot, control engineers can easily design controllers that suppress disturbances and account for perturbations of the controlled object with frequency response data alone. There is no need to construct the transfer function models of the controlled object and of the disturbance. Experimental results for a track-following control in an HDD show the utility of the method. © 2011 IEEE.


Cho R.,University of Pittsburgh | Wu W.,Carnegie Mellon University
Frontiers in Psychiatry | Year: 2013

Recent work on the mechanisms underlying auditory verbal hallucination (AVH) has been heavily informed by self-monitoring accounts that postulate defects in an internal monitoring mechanism as the basis of AVH. A more neglected alternative is an account focusing on defects in auditory processing, namely a spontaneous activation account of auditory activity underlying AVH. Science is often aided by putting theories in competition. Accordingly, a discussion that systematically contrasts the two models of AVH can generate sharper questions that will lead to new avenues of investigation. In this paper, we provide such a theoretical discussion of the two models, drawing strong contrasts between them. We identify a set of challenges for the self-monitoring account and argue that the spontaneous activation account has much in favor of it and should be the default account. Our theoretical overview leads to new questions and issues regarding the explanation of AVH as a subjective phenomenon and its neural basis. Accordingly, we suggest a set of experimental strategies to dissect the underlying mechanisms of AVH in light of the two competing models. We shall contrast two proposed mechanisms of auditory verbal hallucinations (AVH): (a) the family of self-monitoring accounts and (b) a less discussed spontaneous activity account. On the former, a monitoring mechanism tracks whether internal episodes such as inner speech are self- or externally generated while on the latter, spontaneous auditory activity is the primary basis of AVH. In one sense, self-monitoring accounts emphasize "top-down" control mechanisms; spontaneous activity accounts emphasize "bottom-up" sensory mechanisms. The aim of this paper is not to provide a comprehensive literature review on AVH as there have been recent reviews (1, 2). Rather, we believe that it remains an open question what mechanisms underlie AVH in schizophrenia, and that by drawing clear contrasts between alternative models, we can identify experimental directions to explain what causes AVH. Self-monitoring accounts have provided much impetus to current theorizing about AVH, but one salient aspect of our discussion is to raise questions as to whether such accounts, as currently formulated, can adequately explain AVH. We believe that there are in fact significant limitations to the account that have largely gone unnoticed. Still, both models we consider might hold, and this requires further empirical investigation. Conceptual and logical analysis, however, will play an important role in aiding empirical work. © 2013 Cho and Wu.


Pikhurko O.,Carnegie Mellon University
European Journal of Combinatorics | Year: 2011

Let Gi be the (unique) 3-graph with 4 vertices and i edges. Razborov [A. Razborov, On 3-hypergraphs with forbidden 4-vertex configurations, SIAM J. Discrete Math. 24 (2010) 946-963] determined asymptotically the minimum size of a 3-graph on n vertices having neither G0 nor G3 as an induced subgraph. Here we obtain the corresponding stability result, determine the extremal function exactly, and describe all extremal hypergraphs for n≥n0. It follows that any sequence of almost extremal hypergraphs converges, which answers in the affirmative a question posed by Razborov. © 2011 Elsevier Ltd.


Guo C.,Fudan University | Guo C.,Carnegie Mellon University | Zhang L.,Fudan University
IEEE Transactions on Image Processing | Year: 2010

Salient areas in natural scenes are generally regarded as areas which the human eye will typically focus on, and finding these areas is the key step in object detection. In computer vision, many models have been proposed to simulate the behavior of eyes such as SaliencyToolBox (STB), Neuromorphic Vision Toolkit (NVT), and others, but they demand high computational cost and computing useful results mostly relies on their choice of parameters. Although some region-based approaches were proposed to reduce the computational complexity of feature maps, these approaches still were not able to work in real time. Recently, a simple and fast approach called spectral residual (SR) was proposed, which uses the SR of the amplitude spectrum to calculate the image's saliency map. However, in our previous work, we pointed out that it is the phase spectrum, not the amplitude spectrum, of an image's Fourier transform that is key to calculating the location of salient areas, and proposed the phase spectrum of Fourier transform (PFT) model. In this paper, we present a quaternion representation of an image which is composed of intensity, color, and motion features. Based on the principle of PFT, a novel multiresolution spatiotemporal saliency detection model called phase spectrum of quaternion Fourier transform (PQFT) is proposed in this paper to calculate the spatiotemporal saliency map of an image by its quaternion representation. Distinct from other models, the added motion dimension allows the phase spectrum to represent spatiotemporal saliency in order to perform attention selection not only for images but also for videos. In addition, the PQFT model can compute the saliency map of an image under various resolutions from coarse to fine. Therefore, the hierarchical selectivity (HS) framework based on the PQFT model is introduced here to construct the tree structure representation of an image.With the help of HS, a model called multiresolution wavelet domain foveation (MWDF) is proposed in this paper to improve coding efficiency in image and video compression. Extensive tests of videos, natural images, and psychological patterns show that the proposed PQFT model is more effective in saliency detection and can predict eye fixations better than other state-of-the-art models in previous literature. Moreover, our model requires low computational cost and, therefore, can work in real time. Additional experiments on image and video compression show that the HS-MWDF model can achieve higher compression rate than the traditional model. © 2009 IEEE.


Thiessen E.D.,Carnegie Mellon University | Yee M.N.,Indiana University
Child Development | Year: 2010

Whereas young children accept words that differ by only a single phoneme as equivalent labels for novel objects, older children do not (J. F. Werker, C. J. Fennell, K. M. Corcoran, & C. L. Stager, 2002). In these experiments, 106 children were exposed to a training regime that has previously been found to facilitate children's use of phonemic contrasts (E. D. Thiessen, 2007). The results indicate that the effect of this training is limited to contexts that are highly similar to children's initial experience with the phonemic contrast, suggesting that early word-form representations are not composed of entirely abstract units such as phonemes or features. Instead, these results are consistent with theories suggesting that children's early word-form representations retain contextual and perceptual features associated with children's prior experience with words. © 2010, Copyright the Author(s). Journal Compilation © 2010, Society for Research in Child Development, Inc.


Reynolds K.A.,RAND Corporation | Helgeson V.S.,Carnegie Mellon University
Annals of Behavioral Medicine | Year: 2011

Background: It is not clear from the literature whether children with diabetes have more psychological difficulties than their peers. Purpose: This study aims to use meta-analysis to determine if children with diabetes differ from children without a chronic illness in a variety of domains reflecting psychological well-being. Method: A meta-analysis was undertaken of 22 studies that compared children with diabetes to a comparison group. Outcomes included depression, anxiety, behavioral problems, and related constructs. Results: Children with diabetes were more likely than comparison groups to experience a variety of psychological difficulties. However, these effects were small to medium in magnitude and were typically smaller among more recent studies and studies with well-matched comparison groups. Conclusions: This meta-analysis suggests that children with diabetes are at slightly elevated risk for psychological difficulties. Future work will need to help identify children at the highest risk, and to identify factors associated with resilience. © 2011 The Society of Behavioral Medicine.


Lu H.-Z.,University of Hong Kong | Yao W.,University of Hong Kong | Xiao D.,Carnegie Mellon University | Shen S.-Q.,University of Hong Kong
Physical Review Letters | Year: 2013

We study the quantum diffusive transport of multivalley massive Dirac cones, where time-reversal symmetry requires opposite spin orientations in inequivalent valleys. We show that the intervalley scattering and intravalley scattering can be distinguished from the quantum conductivity that corrects the semiclassical Drude conductivity, due to their distinct symmetries and localization trends. In immediate practice, it allows transport measurements to estimate the intervalley scattering rate in hole-doped monolayers of group-VI transition metal dichalcogenides (e.g., molybdenum dichalcogenides and tungsten dichalcogenides), an ideal class of materials for valleytronics applications. The results can be generalized to a large class of multivalley massive Dirac systems with spin-valley coupling and time-reversal symmetry. © 2013 American Physical Society.


Jin R.,Carnegie Mellon University
Angewandte Chemie - International Edition | Year: 2010

(Figure Presented) Seeing double: Nanoparticle clusters (dimers, trimers, etc.) have long been pursued as enhancers in surfaceenhanced Raman spectroscopy research. A recent report presents an elegant approach for the high-yielding fabrication of dimers of silver nanospheres from nanocubes by controlled chemical etching. These nanoparticle dimers are capable of strongly enhancing Raman signals of surface adsorbates (see picture). © 2010 Wiley-VCH Verlag GmbH & Co. KCaA,.


Rohrer G.S.,Carnegie Mellon University
Journal of the American Ceramic Society | Year: 2011

Recently developed techniques to measure the structure of interfacial networks in three dimensions have the potential to revolutionize our ability to control the microstructures of polycrystals and interface-dominated materials properties. This paper reviews recent findings from two- and three-dimensional orientation mapping studies. The observations confirm a strong inverse correlation between the relative energies of grain boundaries and the frequency with which they occur in microstructures. The observations also show that during microstructure evolution, relatively higher energy grain boundaries are more likely to be shrinking while lower energy interfaces are more likely to be growing. These processes can lead to a steady-state distribution of grain boundaries that is influenced as much by the relative grain-boundary energies as by the exact processing conditions. Recent findings and emerging opportunities for grain-boundary characterization are reviewed in the final section of the paper. © 2011 The American Ceramic Society.


Willis K.D.D.,Carnegie Mellon University | Wilson A.D.,Microsoft
ACM Transactions on Graphics | Year: 2013

We introduce InfraStructs, material-based tags that embed informa- tion inside digitally fabricated objects for imaging in the Terahertz region. Terahertz imaging can safely penetrate many common ma- terials, opening up new possibilities for encoding hidden informa- tion as part of the fabrication process. We outline the design, fabri- cation, imaging, and data processing steps to fabricate information inside physical objects. Prototype tag designs are presented for lo- cation encoding, pose estimation, object identification, data storage, and authentication. We provide detailed analysis of the constraints and performance considerations for designing InfraStruct tags. Fu- ture application scenarios range from production line inventory, to customized game accessories, to mobile robotics. Copyright © ACM. Copyright © ACM 2013.


Procaccia A.D.,Carnegie Mellon University
Communications of the ACM | Year: 2013

Experts suggest that computer scientists need to play a key role in proper allocation of limited resources between unlimited wants and demands. Experts state that such fair division is considered a significant subfield of microeconomic theory. Experts illustrate the example of cake cutting to reveal the significance of fair allocation of resources among all concerned parties. Cake cutting is a useful metaphor for the more formal-sounding task of allocating a heterogeneous divisible good among multiple players with different preferences. The study of fair cake-cutting algorithms has originated with Steinhaus, Knaster, and Banach in Poland during World War II and has attracted mathematicians, economists and political scientists over a period of time. Reasoning about the complexity of cake-cutting algorithms requires a model specifying what such an algorithm can do.


Wagner S.J.,University of Maine, United States | Rubin E.S.,Carnegie Mellon University
Renewable Energy | Year: 2013

Solar energy is an attractive renewable energy source because the sun's energy is plentiful and carbon-free. However, solar energy is intermittent and not suitable for base load electricity generation without an energy backup system. Concentrated solar power (CSP) is unique among other renewable energy options because it can approach base load generation with molten salt thermal energy storage (TES). This paper describes the development of an engineering economic model that directly compares the performance, cost, and profit of a 110-MW parabolic trough CSP plant operating with a TES system, natural gas-fired backup system, and no backup system. Model results are presented for 0-12h backup capacities with and without current U.S. subsidies. TES increased the annual capacity factor from around 30% with no backup to up to 55% with 12h of storage when the solar field area was selected to provide the lowest levelized cost of energy (LCOE). Using TES instead of a natural gas-fired heat transfer fluid heater (NG) increased total plant capital costs but decreased annual operation and maintenance costs. These three effects led to an increase in the LCOE for PT plants with TES and NG backup compared with no backup. LCOE increased with increasing backup capacity for plants with TES and NG backup. For small backup capacities (1-4h), plants with TES had slightly lower LCOE values than plants with NG backup. For larger backup capacities (5-12h), plants with TES had slightly higher LCOE values than plants with NG backup. At these costs, current U.S. federal tax incentives were not sufficient to make PT profitable in a market with variable electricity pricing. Current U.S. incentives combined with a fixed electricity price of $200/MWh made PT plants with larger backup capacities more profitable than PT plants with no backup or with smaller backup capacities. In the absence of incentives, a carbon price of $100-$160/tonne CO2eq would be required for these PT plants to compete with new coal-fired power plants in the U.S. If the long-term goal is to increase renewable base load electricity generation, additional incentives are needed to encourage new CSP plants to use thermal energy storage in the U.S. © 2012 Elsevier Ltd.


Brown R.D.,Carnegie Mellon University
Digital Investigation | Year: 2012

This paper presents a trainable open-source utility to extract text from arbitrary data files and disk images which uses language models to automatically detect character encodings prior to extracting strings and for automatic language identification and filtering of non-textual strings after extraction. With a test set containing 923 languages, consisting of strings of at most 65 characters, an overall language identification error rate of less than 0.4% is achieved. False-alarm rates on random data are 0.34% when filtering thresholds are set for high recall and 0.012% when set for high precision, with corresponding miss rates of 0.002% and 0.009% in running text. © 2012 Dykstra & Sherman. Published by Elsevier Ltd. All rights reserved.


Seo T.,Yeungnam University | Sitti M.,Carnegie Mellon University
IEEE/ASME Transactions on Mechatronics | Year: 2013

This paper proposes an underactuated modular climbing robot with flat dry elastomer adhesives. This robot is designed to achieve high speed, high payload, and dexterous motions that are typical drawbacks of previous climbing robots. Each module is designed as a tread-wheeled mechanism to simultaneously realize high speed and high adhesive force. Two modules are connected by compliant joints, which induce a positive preload on the front wheels resulting in stable climbing and high payload capacity. Compliant joints also help the robot to perform various transitions. An active tail is adopted to regulate the preload of the second module. Force transfer equations are derived and stable operating conditions are verified. The stiffness coefficients of the compliant joints and the active tail force are determined optimally to satisfy the constraints of stable operation. The prototype two-module robot achieves 6-cm/\hbox{s} speed and 500-g payload capacity on vertical surfaces. The abilities of flat surface locomotion, internal, external, and thin-wall transitions, and overcoming various sized obstacles are validated through experiment. The principle of joint compliance can be adopted in other climbing robots to enhance their stability and transition capability. © 1996-2012 IEEE.


Xu X.,University of Washington | Yao W.,University of Hong Kong | Xiao D.,Carnegie Mellon University | Heinz T.F.,Columbia University
Nature Physics | Year: 2014

The recent emergence of two-dimensional layered materials-in particular the transition metal dichalcogenides-provides a new laboratory for exploring the internal quantum degrees of freedom of electrons and their potential for new electronics. These degrees of freedom are the real electron spin, the layer pseudospin, and the valley pseudospin. New methods for the quantum control of the spin and these pseudospins arise from the existence of Berry phase-related physical properties and strong spin-orbit coupling. The former leads to the versatile control of the valley pseudospin, whereas the latter gives rise to an interplay between the spin and the pseudospins. Here, we provide a brief review of both theoretical and experimental advances in this field. © 2014 Macmillan Publishers Limited.


De La Torre F.,Carnegie Mellon University
ACM Transactions on Graphics | Year: 2011

Linear models, particularly those based on principal component analysis (PCA), have been used successfully on a broad range of human face-related applications. Although PCA models achieve high compression, they have not been widely used for animation in a production environment because their bases lack a semantic interpretation. Their parameters are not an intuitive set for animators to work with. In this paper we present a linear face modelling approach that generalises to unseen data better than the traditional holistic approach while also allowing click-and-drag interaction for animation. Our model is composed of a collection of PCA sub-models that are independently trained but share boundaries. Boundary consistency and user-given constraints are enforced in a soft least mean squares sense to give flexibility to the model while maintaining coherence. Our results show that the region-based model generalizes better than its holistic counterpart when describing previously unseen motion capture data from multiple subjects. The decomposition of the face into several regions, which we determine automatically from training data, gives the user localised manipulation control. This feature allows to use the model for face posing and animation in an intuitive style. © 2011 ACM.


Deserno M.,Carnegie Mellon University
Chemistry and Physics of Lipids | Year: 2015

A fluid lipid membrane transmits stresses and torques that are fully determined by its geometry. They can be described by a stress- and torque-tensor, respectively, which yield the force or torque per length through any curve drawn on the membrane's surface. In the absence of external forces or torques the surface divergence of these tensors vanishes, revealing them as conserved quantities of the underlying Euler-Lagrange equation for the membrane's shape. This review provides a comprehensive introduction into these concepts without assuming the reader's familiarity with differential geometry, which instead will be developed as needed, relying on little more than vector calculus. The Helfrich Hamiltonian is then introduced and discussed in some depth. By expressing the quest for the energy-minimizing shape as a functional variation problem subject to geometric constraints, as proposed by Guven (2004), stress- and torque-tensors naturally emerge, and their connection to the shape equation becomes evident. How to reason with both tensors is then illustrated with a number of simple examples, after which this review concludes with four more sophisticated applications: boundary conditions for adhering membranes, corrections to the classical micropipette aspiration equation, membrane buckling, and membrane mediated interactions. © 2014 Elsevier Ireland Ltd. All rights reserved.


Guruswami V.,Carnegie Mellon University
Proceedings of the Annual IEEE Conference on Computational Complexity | Year: 2011

Folded Reed-Solomon codes are an explicit family of codes that achieve the optimal trade-off between rate and error-correction capability: specifically, for any ε > 0, the author and Rudra (2006, 08) presented an n O(1ε) time algorithm to list decode appropriate folded RS codes of rate R from a fraction 1 - R - ε of errors. The algorithm is based on multivariate polynomial interpolation and root-finding over extension fields. It was noted by Vadhan that interpolating a linear polynomial suffices if one settles for a smaller decoding radius (but still enough for a statement of the above form). Here we give a simple linear-algebra based analysis of this variant that eliminates the need for the computationally expensive root-finding step over extension fields (and indeed any mention of extension fields). The entire list decoding algorithm is linear-algebraic, solving one linear system for the interpolation step, and another linear system to find a small subspace of candidate solutions. Except for the step of pruning this subspace, the algorithm can be implemented to run in quadratic time. The theoretical drawback of folded RS codes are that both the decoding complexity and proven worst-case list-size bound are nΩ(1/ ε). By combining the above idea with a pseudorandom subset of all polynomials as messages, we get a Monte Carlo construction achieving a list size bound of O(1/ε2) which is quite close to the existential O(1/ε) bound (however, the decoding complexity remains n Ω(1/ε)). Our work highlights that constructing an explicit subspace-evasive subset that has small intersection with low-dimensional subspaces - an interesting problem in pseudorandomness in its own right - could lead to explicit codes with better list-decoding guarantees. © 2011 IEEE.


Zhu R.,Carnegie Mellon University
IEEE International Conference on Intelligent Robots and Systems | Year: 2012

We describe the software components of a robotics system designed to autonomously grasp objects and perform dexterous manipulation tasks with only high-level supervision. The system is centered on the tight integration of several core functionalities, including perception, planning and control, with the logical structuring of tasks driven by a Behavior Tree architecture. The advantage of the implementation is to reduce the execution time while integrating advanced algorithms for autonomous manipulation. We describe our approach to 3-D perception, real-time planning, force compliant motions, and audio processing. Performance results for object grasping and complex manipulation tasks of in-house tests and of an independent evaluation team are presented. © 2012 IEEE.


Gurtin M.E.,Carnegie Mellon University | Reddy B.D.,University of Cape Town
Journal of the Mechanics and Physics of Solids | Year: 2014

This paper develops a theory of rate-independent single-crystal plasticity at small length scales. The theory is thermodynamically consistent, and makes provision for power expenditures resulting from vector and scalar microscopic stresses respectively conjugate to slip rates and their tangential gradients on the individual slip systems. Scalar generalized accumulated slips form the basis for a new hardening relation, which takes account of self- and latent-hardening. The resulting initial-boundary value problem is placed in a variational setting in the form of a global variational inequality. © 2014 Elsevier Ltd. All rights reserved.


Bar-Joseph Z.,Carnegie Mellon University
BMC genomics | Year: 2013

Studies of gene regulation often utilize genome-wide predictions of transcription factor (TF) binding sites. Most existing prediction methods are based on sequence information alone, ignoring biological contexts such as developmental stages and tissue types. Experimental methods to study in vivo binding, including ChIP-chip and ChIP-seq, can only study one transcription factor in a single cell type and under a specific condition in each experiment, and therefore cannot scale to determine the full set of regulatory interactions in mammalian transcriptional regulatory networks. We developed a new computational approach, PIPES, for predicting tissue-specific TF binding. PIPES integrates in vitro protein binding microarrays (PBMs), sequence conservation and tissue-specific epigenetic (DNase I hypersensitivity) information. We demonstrate that PIPES improves over existing methods on distinguishing between in vivo bound and unbound sequences using ChIP-seq data for 11 mouse TFs. In addition, our predictions are in good agreement with current knowledge of tissue-specific TF regulation. We provide a systematic map of computationally predicted tissue-specific binding targets for 284 mouse TFs across 55 tissue/cell types. Such comprehensive resource is useful for researchers studying gene regulation.


Zhu J.-G.,Carnegie Mellon University
IEEE Transactions on Magnetics | Year: 2014

In this paper, we present a micromagnetic modeling study on the switching property of segmented grains assisted by circularly polarized magnetic field at microwave frequencies and the actual recording process. This paper provides insightful understanding of the switching dependence on the Gilbert damping constant and the optimization for suppressing the impact of possible damping constant distribution. The recording processes with ac field generated by a spin torque oscillator on the media consisting of segmented grains are also investigated. In this microwave assisted magnetic recording (MAMR) process, the segmentation of the grains with exchange breaking layers enables the best utilization of the ac field. The MAMR simulations show significant signal-to-noise ratio (SNR) gain over today's perpendicular recording for medium for media with grain pitches <8 nm. Grain-pitch limited SNR performance can be achieved for medium grain pitches as small as 5 nm while achieving sufficient thermal stability simultaneously. At 4.5 nm grain pitch, MAMR shows significantly higher SNR than that predicted by corresponding HAMR simulations. It is concluded that MAMR should enable areal density capability of 4 Tb/in 2 and beyond and its SNR performance is superior to HAMR at all densities. © 2014 IEEE.


Salerno K.M.,Johns Hopkins University | Maloney C.E.,Carnegie Mellon University | Robbins M.O.,Johns Hopkins University
Physical Review Letters | Year: 2012

Simulations are used to determine the effect of inertia on athermal shear of amorphous two-dimensional solids. In the quasistatic limit, shear occurs through a series of rapid avalanches. The distribution of avalanches is analyzed using finite-size scaling with thousands to millions of disks. Inertia takes the system to a new underdamped universality class rather than driving the system away from criticality as previously thought. Scaling exponents are determined for the underdamped and overdamped limits and a critical damping that separates the two regimes. Systems are in the overdamped universality class even when most vibrational modes are underdamped. © 2012 American Physical Society.


Frieze A.,Carnegie Mellon University
Electronic Journal of Combinatorics | Year: 2010

In the random hypergraph H = Hn,p;3 each possible triple appears indepen- dently with probability p. A loose Hamilton cycle can be described as a sequence of edges {xi, yi, xi+1} for i = 1, 2,.n/2 where x1, x2,.,xn/2, y1, y2,.,yn/2 are all distinct. We prove that there exists an absolute constant K > 0 such that if p ≥ K log n/n2 then lim/n→∞/4n Pr(Hn,p;3 contains a loose Hamilton cycle) = 1.


Reddy R.,Carnegie Mellon University
Communications of the ACM | Year: 2014

Experts provide their collective historical perspective on the advances in the field of speech recognition. They limit the scope to discussing the missing science of speech recognition 40 years ago and the advances that have contributed to overcoming some of the most challenging problems. The insights gained from the speech recognition advances over the past 40 years are explored, originating from generations of Carnegie Mellon University's R&D. Several major achievements over the years have proven to work well in practice for leading industry speech recognition systems ranging from Apple to Microsoft. Improved ability of devices to handle relatively unrestricted multimodal dialogues is being realized with the adoption of speech recognition used in Apple, Google, and Microsoft products.


Sacchini J.L.,Carnegie Mellon University
Proceedings - Symposium on Logic in Computer Science | Year: 2013

Productivity of core cursive definitions is an essential property in proof assistants since it ensures logical consistency and decidability of type checking. Type-based mechanisms for ensuring productivity use types annotated with size information to track the number of elements produced in core cursive definitions. In this paper, we propose an extension of the Calculus of Constructions-the theory underlying the Coq proof assistant-with a type-based criterion for ensuring productivity of stream definitions. We prove strong normalization and logical consistency. Furthermore, we define an algorithm for inferring size annotations in types. These results can be easily extended to handle general co inductive types. © 2013 IEEE.


Haviland A.M.,Carnegie Mellon University
Medical care | Year: 2012

To examine racial/ethnic differences in Medicare beneficiary experiences with Medicare Part D prescription drug (PD) coverage. 2008 Consumer Assessment of Health Care Providers and Systems survey of U.S. Medicare beneficiaries. Surveys were administered by mail with phone follow-up to a nationally representative sample (61% response rate). This study examines 201,496 beneficiaries of age 65 and older with PD coverage [6% Hispanic, 7% non-Hispanic Black, 3% non-Hispanic Asian or Pacific Islander (API)]. Key variables are self-reported race/ethnicity and Consumer Assessment of Health Care Providers and Systems getting information and needed PDs measures. We generated weighted case-mix adjusted means for 4 racial/ethnic groups and for Hispanics separately by English-language or Spanish-language preference status. We calculated within-plan disparities using a linear mixed-effect model, with fixed effects for race/ethnicity, coverage type and case-mix variables, and random effects for contract and contract by race/ethnicity interactions. Disparities for Hispanic, Black, and API beneficiaries on obtaining needed PDs and information regarding coverage range from -2 to -11 points (0-100 scale) relative to non-Hispanic Whites, with the greatest disparities observed for Spanish-preferring Hispanics and API beneficiaries, especially those with low income. There is wide variation in disparities across contracts, and contracts with the largest disparities for Hispanics have higher proportions of beneficiaries with lower education and income. Quality improvement efforts may be needed to reduce racial/ethnic disparities in beneficiary experience with PD coverage. Cultural, language, and health literacy barriers in navigating Medicare's Part D program may partially explain the observed disparities.


Whalen D.J.,Carnegie Mellon University | Fryer C.L.,Los Alamos National Laboratory
Astrophysical Journal Letters | Year: 2012

The existence of 109 M ⊙ black holes (BHs) in massive galaxies by z ∼ 7 is one of the great unsolved mysteries in cosmological structure formation. One theory argues that they originate from the BHs of Pop III stars at z ∼ 20 and then accrete at the Eddington limit down to the epoch of reionization, which requires that they have constant access to rich supplies of fuel. Because early numerical simulations suggested that Pop III stars were ≳100 M ⊙, the supermassive black hole (SMBH) seeds considered up to now were 100-300 M ⊙. However, there is a growing numerical and observational consensus that some Pop III stars were tens of solar masses, not hundreds, and that 20-40 M ⊙ BHs may have been much more plentiful at high redshift. However, we find that natal kicks imparted to 20-40 M ⊙ Pop III BHs during formation eject them from their halos and hence their fuel supply, precluding them from Eddington-limit growth. Consequently, SMBHs are far less likely to form from low-mass Pop III stars than from very massive ones. © © 2012. The American Astronomical Society. All rights reserved.


Biegler L.T.,Carnegie Mellon University
Chemie-Ingenieur-Technik | Year: 2014

Systematic optimization strategies are essential in process design, operations and control. This study provides a concise overview of recent advances in process optimization, with a particular focus on nonlinear programming. Here, derivativefree, SQP, reduced gradient, and interior point methods are described. Moreover, state of the art problem formulations and modeling environments are described, and contrasted with popular process simulation tools. Finally research directions are identified for future trends in multi-scale optimization within large-scale equation-oriented optimization environments. © 2014 Wiley-VCH Verlag GmbH and Co. KGaA, Weinheim.


Jin R.,Carnegie Mellon University | Nobusada K.,Japan Institute for Molecular Science | Nobusada K.,Kyoto University
Nano Research | Year: 2014

The recent success in the synthesis and total structure determination of atomically precise gold nanoparticles has provided exciting opportunities for fundamental studies as well as the development of new applications. These unique nanoparticles are of molecular purity and possess well defined formulas (i.e., specific numbers of metal atoms and ligands), resembling organic compounds. Crystallization of such molecularly pure nanoparticles into macroscopic single crystals allows for the determination of total structures of nanoparticles (i.e., the arrangement of metal core atoms and surface ligands) by X-ray crystallography. In this perspective article, we summarize recent efforts in doping and alloying gold nanoparticles with other metals, including Pd, Pt, Ag and Cu. With atomically precise gold nanoparticles, a specific number of foreign atoms (e.g., Pd, Pt) can be incorporated into the gold core, whereas a range of Ag and Cu substitutions is observed but, interestingly, the total number of metal atoms in the homogold nanoparticle is preserved. The heteroatom substitution of gold nanoparticles allows one to probe the optical, structural, and electronic properties truly at the single-atom level, and thus provides a wealth of information for understanding the intriguing properties of this new class of nanomaterials. [Figure not available: see fulltext.] © 2014 Tsinghua University Press and Springer-Verlag Berlin Heidelberg.


Bettinger C.J.,Carnegie Mellon University
Macromolecular Bioscience | Year: 2011

Synthetic biomaterials serve as a cornerstone in the development of clinically focused regenerative medicine therapies that aim to reduce suffering and prolong life. Recent improvements in biodegradable elastomeric materials utilize natural extracellular matrix proteins as inspiration to yield a new class of materials with superior degradation kinetics, desirable biocompatibility profiles, and mechanical properties that closely match those of soft tissues. This review describes several classes of synthetic biodegradable elastomers and associated fabrication techniques that are relevant to scaffold development. The application of these materials to select tissue engineering models is also discussed. Synthetic biodegradable elastomers use natural extracellular matrix proteins as inspiration to yield new classes of materials with superior properties for soft tissue engineering. This review describes recent advancements in synthetic biodegradable elastomers as well as emerging applications in regenerative medicine. © 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.


Mechanotransduction of sensory neurons is of great interest to the scientific community, especially in areas such as pain, neurobiology, cardiovascular homeostasis and mechanobiology. We describe a method to investigate stretch-activated mechanotransduction in sensory nerves through subcellular stimulation. The method imposes localized mechanical stimulation through indentation of an elastomeric substrate and combines this mechanical stimulation with whole-cell patch clamp recording of the electrical response to single-nerve stretching. One significant advantage here is that the neurites are stretched with limited physical contact beyond their attachment to the polymer. When we imposed specific mechanical stimulation through the substrate, the stretched neurite fired and an action potential response was recorded. In addition, complementary protocols to control the molecules at the cell-substrate interface are presented. These techniques provide an opportunity to probe neurosensory mechanotransduction with a defined substrate, whose physical and molecular context can be modified to mimic physiologically relevant conditions. The entire process from fabrication to cellular recording takes 5 to 6 d.


Balcan M.-F.,Georgia Institute of Technology | Blum A.,Carnegie Mellon University
Journal of the ACM | Year: 2010

Supervised learning-that is, learning from labeled examples-is an area of Machine Learning that has reached substantial maturity. It has generated general-purpose and practically successful algorithms and the foundations are quite well understood and captured by theoretical frameworks such as the PAC-learning model and the Statistical Learning theory framework. However, for many contemporary practical problems such as classifying web pages or detecting spam, there is often additional information available in the form of unlabeled data, which is often much cheaper and more plentiful than labeled data. As a consequence, there has recently been substantial interest in semi-supervised learningusing unlabeled data together with labeled datasince any useful information that reduces the amount of labeled data needed can be a significant benefit. Several techniques have been developed for doing this, along with experimental results on a variety of different learning problems. Unfortunately, the standard learning frameworks for reasoning about supervised learning do not capture the key aspects and the assumptions underlying these semi-supervised learning methods. In this article, we describe an augmented version of the PAC model designed for semi-supervised learning, that can be used to reason about many of the different approaches taken over the past decade in the Machine Learning community. This model provides a unified framework for analyzing when and why unlabeled data can help, in which one can analyze both sample-complexity and algorithmic issues. The model can be viewed as an extension of the standard PAC model where, in addition to a concept class C, one also proposes a compatibility notion: a type of compatibility that one believes the target concept should have with the underlying distribution of data. Unlabeled data is then potentially helpful in this setting because it allows one to estimate compatibility over the space of hypotheses, and to reduce the size of the search space from the whole set of hypotheses C down to those that, according to one's assumptions, are a-priori reasonable with respect to the distribution. As we show, many of the assumptions underlying existing semi-supervised learning algorithms can be formulated in this framework. After proposing the model, we then analyze sample-complexity issues in this setting: that is, how much of each type of data one should expect to need in order to learn well, and what the key quantities are that these numbers depend on. We also consider the algorithmic question of how to efficiently optimize for natural classes and compatibility notions, and provide several algorithmic results including an improved bound for Co-Training with linear separators when the distribution satisfies independence given the label. © 2010 ACM.


Platzer A.,Carnegie Mellon University
Logical Methods in Computer Science | Year: 2012

The biggest challenge in hybrid systems verification is the handling of differential equations. Because computable closed-form solutions only exist for very simple differential equations, proof certificates have been proposed for more scalable verification. Search procedures for these proof certificates are still rather ad-hoc, though, because the problem structure is only understood poorly. We investigate differential invariants, which define an induction principle for differential equations and which can be checked for invariance along a differential equation just by using their differential structure, without having to solve them. We study the structural properties of differential invariants. To analyze trade-offs for proof search complexity, we identify more than a dozen relations between several classes of differential invariants and compare their deductive power. As our main results, we analyze the deductive power of differential cuts and the deductive power of differential invariants with auxiliary differential variables. We refute the differential cut elimination hypothesis and show that, unlike standard cuts, differential cuts are fundamental proof principles that strictly increase the deductive power. We also prove that the deductive power increases further when adding auxiliary differential variables to the dynamics.


Platzer A.,Carnegie Mellon University
Logical Methods in Computer Science | Year: 2012

We address a fundamental mismatch between the combinations of dynamics that occur in cyber-physical systems and the limited kinds of dynamics supported in analysis. Modern applications combine communication, computation, and control. They may even form dynamic distributed networks, where neither structure nor dimension stay the same while the system follows hybrid dynamics, i.e., mixed discrete and continuous dynamics. We provide the logical foundations for closing this analytic gap. We develop a formal model for distributed hybrid systems. It combines quantified differential equations with quantified assignments and dynamic dimensionality-changes. We introduce a dynamic logic for verifying distributed hybrid systems and present a proof calculus for this logic. This is the first formal verification approach for distributed hybrid systems. We prove that our calculus is a sound and complete axiomatization of the behavior of distributed hybrid systems relative to quantified differential equations. In our calculus we have proven collision freedom in distributed car control even when an unbounded number of new cars may appear dynamically on the road.


De La Torre F.,Carnegie Mellon University
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2012

Over the last century, Component Analysis (CA) methods such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Canonical Correlation Analysis (CCA), Locality Preserving Projections (LPP), and Spectral Clustering (SC) have been extensively used as a feature extraction step for modeling, classification, visualization, and clustering. CA techniques are appealing because many can be formulated as eigen-problems, offering great potential for learning linear and nonlinear representations of data in closed-form. However, the eigen-formulation often conceals important analytic and computational drawbacks of CA techniques, such as solving generalized eigen-problems with rank deficient matrices (e.g., small sample size problem), lacking intuitive interpretation of normalization factors, and understanding commonalities and differences between CA methods. This paper proposes a unified least-squares framework to formulate many CA methods. We show how PCA, LDA, CCA, LPP, SC, and its kernel and regularized extensions correspond to a particular instance of least-squares weighted kernel reduced rank regression (LS-WKRRR). The LS-WKRRR formulation of CA methods has several benefits: 1) provides a clean connection between many CA techniques and an intuitive framework to understand normalization factors; 2) yields efficient numerical schemes to solve CA techniques; 3) overcomes the small sample size problem; 4) provides a framework to easily extend CA methods. We derive weighted generalizations of PCA, LDA, SC, and CCA, and several new CA techniques. © 2012 IEEE.


Acharya A.,Carnegie Mellon University
Journal of the Mechanics and Physics of Solids | Year: 2010

Nonsingular, stressed, dislocation (wall) profiles are shown to be 1-d equilibria of a non-equilibrium theory of Field Dislocation Mechanics (FDM). It is also shown that such equilibrium profiles corresponding to a given level of load cannot generally serve as a travelling wave profile of the governing equation for other values of nearby constant load; however, one case of soft loading with a special form of the dislocation velocity law is demonstrated to have no 'Peierls barrier' in this sense. The analysis is facilitated by the formulation of a 1-d, scalar, time-dependent, Hamilton-Jacobi equation as an exact special case of the full 3-d FDM theory accounting for non-convex elastic energy, small, Nye-tensor-dependent core energy, and possibly an energy contribution based on incompatible slip. Relevant nonlinear stability questions, including that of nucleation, are formulated in a non-equilibrium setting. Elementary averaging ideas show a singular perturbation structure in the evolution of the (unsymmetric) macroscopic plastic distortion, thus pointing to the possibility of predicting generally rate-insensitive slow response constrained to a tensorial 'yield' surface, while allowing fast excursions off it, even though only simple kinetic assumptions are employed in the microscopic FDM theory. The emergent small viscosity on averaging that serves as the small parameter for the perturbation structure is a robust, almost-geometric consequence of large gradients of slip in the dislocation core and the persistent presence of a large number of dislocations in the averaging volume. In the simplest approximation, the macroscopic yield criterion displays anisotropy based on the microscopic dislocation line and Burgers vector distribution, a dependence on the Laplacian of the incompatible slip tensor and a nonlocal term related to a Stokes-Helmholtz-curl projection of an 'internal stress' derived from the incompatible slip energy. © 2010.


Ortiz E.G.,University of Central Florida | Becker B.C.,Carnegie Mellon University
Computer Vision and Image Understanding | Year: 2014

With millions of users and billions of photos, web-scale face recognition is a challenging task that demands speed, accuracy, and scalability. Most current approaches do not address and do not scale well to Internet-sized scenarios such as tagging friends or finding celebrities. Focusing on web-scale face identification, we gather an 800,000 face dataset from the Facebook social network that models real-world situations where specific faces must be recognized and unknown identities rejected. We propose a novel Linearly Approximated Sparse Representation-based Classification (LASRC) algorithm that uses linear regression to perform sample selection for ℓ1- minimization, thus harnessing the speed of least-squares and the robustness of sparse solutions such as SRC. Our efficient LASRC algorithm achieves comparable performance to SRC with a 100-250 times speedup and exhibits similar recall to SVMs with much faster training. Extensive tests demonstrate our proposed approach is competitive on pair-matching verification tasks and outperforms current state-of-the-art algorithms on open-universe identification in uncontrolled, web-scale scenarios. © 2013 Elsevier Inc. All rights reserved.


Ali S.,Carnegie Mellon University | Shah M.,University of Central Florida
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2010

We propose a set of kinematic features that are derived from the optical flow for human action recognition in videos. The set of kinematic features includes divergence, vorticity, symmetric and antisymmetric flow fields, second and third principal invariants of flow gradient and rate of strain tensor, and third principal invariant of rate of rotation tensor. Each kinematic feature, when computed from the optical flow of a sequence of images, gives rise to a spatiotemporal pattern. It is then assumed that the representative dynamics of the optical flow are captured by these spatiotemporal patterns in the form of dominant kinematic trends or kinematic modes. These kinematic modes are computed by performing Principal Component Analysis (PCA) on the spatiotemporal volumes of the kinematic features. For classification, we propose the use of multiple instance learning (MIL) in which each action video is represented by a bag of kinematic modes. Each video is then embedded into a kinematic-mode-based feature space and the coordinates of the video in that space are used for classification using the nearest neighbor algorithm. The qualitative and quantitative results are reported on the benchmark data sets. © 2010 IEEE.


Metwally A.,Google | Faloutsos C.,Carnegie Mellon University
Proceedings of the VLDB Endowment | Year: 2012

This work proposes V-SMART-Join, a scalable MapReducebased framework for discovering all pairs of similar entities. The V-SMART-Join framework is applicable to sets, multisets, and vectors. V-SMART-Join is motivated by the observed skew in the underlying distributions of Internet traffic, and is a family of 2-stage algorithms, where the first stage computes and joins the partial results, and the second stage computes the similarity exactly for all candidate pairs. The V-SMART-Join algorithms are very efficient and scalable in the number of entities, as well as their cardinalities. They were up to 30 times faster than the state of the art algorithm, VCL, when compared on a real dataset of a small size. We also established the scalability of the proposed algorithms by running them on a dataset of a realistic size, on which VCL never succeeded to finish. Experiments were run using real datasets of IPs and cookies, where each IP is represented as a multiset of cookies, and the goal is to discover similar IPs to identify Internet proxies. © 2012 VLDB Endowment.


Neill D.B.,Carnegie Mellon University
Statistics in Medicine | Year: 2011

The multivariate Bayesian scan statistic (MBSS) is a recently proposed, general framework for event detection and characterization in multivariate space-time data. MBSS integrates prior information and observations from multiple data streams in a Bayesian framework, computing the posterior probability of each type of event in each space-time region. MBSS has been shown to have many advantages over previous event detection approaches, including improved timeliness and accuracy of detection, easy interpretation and visualization of results, and the ability to model and accurately differentiate between multiple event types. This work extends the MBSS framework to enable detection and visualization of irregularly shaped clusters in multivariate data, by defining a hierarchical prior over all subsets of locations. While a naive search over the exponentially many subsets would be computationally infeasible, we demonstrate that the total posterior probability that each location has been affected can be efficiently computed, enabling rapid detection and visualization of irregular clusters. We compare the run time and detection power of this 'Fast Subset Sums' method to our original MBSS approach (assuming a uniform prior over circular regions) on semi-synthetic outbreaks injected into real-world Emergency Department data from Allegheny County, Pennsylvania. We demonstrate substantial improvements in spatial accuracy and timeliness of detection, while maintaining the scalability and fast run time of the original MBSS method. © 2011 John Wiley & Sons, Ltd.


Sah S.,Georgetown University | Sah S.,Harvard University | Loewenstein G.,Carnegie Mellon University
Psychological Science | Year: 2014

Professionals face conflicts of interest when they have a personal interest in giving biased advice. Mandatory disclosure-informing consumers of the conflict-is a widely adopted strategy in numerous professions, such as medicine, finance, and accounting. Prior research has shown, however, that such disclosures have little impact on consumer behavior, and can backfire by leading advisors to give even more biased advice. We present results from three experiments with real monetary stakes. These results show that, although disclosure has generally been found to be ineffective for dealing with unavoidable conflicts of interest, it can be beneficial when providers have the ability to avoid conflicts. Mandatory and voluntary disclosure can deter advisors from accepting conflicts of interest so that they have nothing to disclose except the absence of conflicts. We propose that people are averse to being viewed as biased, and that policies designed to activate reputational and ethical concerns will motivate advisors to avoid conflicts of interest. © The Author(s) 2013.


Woolford Jr. J.L.,Carnegie Mellon University | Baserga S.J.,Yale University
Genetics | Year: 2013

Ribosomes are highly conserved ribonucleoprotein nanomachines that translate information in the genome to create the proteome in all cells. In yeast these complex particles contain four RNAs (.5400 nucleotides) and 79 different proteins. During the past 25 years, studies in yeast have led the way to understanding how these molecules are assembled into ribosomes in vivo. Assembly begins with transcription of ribosomal RNA in the nucleolus, where the RNA then undergoes complex pathways of folding, coupled with nucleotide modification, removal of spacer sequences, and binding to ribosomal proteins. More than 200 assembly factors and 76 small nucleolar RNAs transiently associate with assembling ribosomes, to enable their accurate and efficient construction. Following export of preribosomes from the nucleus to the cytoplasm, they undergo final stages of maturation before entering the pool of functioning ribosomes. Elaborate mechanisms exist to monitor the formation of correct structural and functional neighborhoods within ribosomes and to destroy preribosomes that fail to assemble properly. Studies of yeast ribosome biogenesis provide useful models for ribosomopathies, diseases in humans that result from failure to properly assemble ribosomes. © 2013 by the Genetics Society of America.


Jin R.,Carnegie Mellon University
Nanoscale | Year: 2010

The scientific study of gold nanoparticles (typically 1-100 nm) has spanned more than 150 years since Faraday's time and will apparently last longer. This review will focus on a special type of ultrasmall (<2 nm) yet robust gold nanoparticles that are protected by thiolates, so-called gold thiolate nanoclusters, denoted as Au n(SR) m (where, n and m represent the number of gold atoms and thiolate ligands, respectively). Despite the past fifteen years' intense work on Au n(SR) m nanoclusters, there is still a tremendous amount of science that is not yet understood, which is mainly hampered by the unavailability of atomically precise Au n(SR) m clusters and by their unknown structures. Nonetheless, recent research advances have opened an avenue to achieving the precise control of Au n(SR) m nanoclusters at the ultimate atomic level. The successful structural determination of Au 102(SPhCOOH) 44 and [Au 25(SCH 2CH 2Ph) 18] q (q = -1, 0) by X-ray crystallography has shed some light on the unique atomic packing structure adopted in these gold thiolate nanoclusters, and has also permitted a precise correlation of their structure with properties, including electronic, optical and magnetic properties. Some exciting research is anticipated to take place in the next few years and may stimulate a long-lasting and wider scientific and technological interest in this special type of Au nanoparticles. © 2010 The Royal Society of Chemistry.


Tonguz O.K.,Carnegie Mellon University
IEEE Communications Magazine | Year: 2011

Traffic congestion in urban areas is an acute problem which is getting worse with the increased urbanization of the world population. The existing approaches to increasing traffic flow in urban areas have proven inefficient as they are expensive and therefore not scalable. It is shown in this article that a biologically inspired new approach could solve some of the fundamental transportation problems via a self-organizing traffic management paradigm. © 2011 IEEE.


Bryant R.E.,Carnegie Mellon University
Computing in Science and Engineering | Year: 2011

Increasingly, scientific computing applications must accumulate and manage massive datasets, as well as perform sophisticated computations over these data. Such applications call for data-intensive scalable computer (DISC) systems, which differ in fundamental ways from existing high-performance computing systems. © 2011 IEEE.


Nagin D.S.,Carnegie Mellon University | Solow R.M.,Massachusetts Institute of Technology | Lum C.,George Mason University
Criminology | Year: 2015

In this article, we join three distinct literatures on crime control-the deterrence literature, the policing literature as it relates to crime control, and the environmental and opportunity perspectives literature. Based on empirical findings and theory from these literatures, we pose a mathematical model of the distribution of criminal opportunities and offender decision making on which of those opportunities to victimize. Criminal opportunities are characterized in terms of the risk of apprehension that attends their victimization. In developing this model, our primary focus is on how police might affect the distribution of criminal opportunities that are attractive to would-be offenders. The theoretical model we pose, however, is generalizable to explain how changes in other relevant target characteristics, such as potential gain, could affect target attractiveness. We demonstrate that the model has important implications for the efficiency and effectiveness of police deployment strategies such as hot spots policing, random patrol, and problem-oriented policing. The theoretical structure also makes clear why the clearance rate is a fundamentally flawed metric of police performance. Future research directions suggested by the theoretical model are discussed. © 2015 American Society of Criminology.


Vaidya V.,Carnegie Mellon University
Physical Review A - Atomic, Molecular, and Optical Physics | Year: 2014

I utilize effective field theory(EFT) techniques to calculate the Casimir torque on a cylindrical gear in the presence of a polarizable but neutral object and present results for the energy and torque as a function of angle for a gear with multiple cogs, as well as for the case of a concentric cylindrical gear. © 2014 American Physical Society.


Stahlke D.,Carnegie Mellon University
Physical Review A - Atomic, Molecular, and Optical Physics | Year: 2014

Quantum states can, in a sense, be thought of as generalizations of classical probability distributions, but are more powerful than probability distributions when used for computation or communication. Quantum speedup therefore requires some feature of quantum states that classical probability distributions lack. One such feature is interference. We quantify interference and show that there can be no quantum speedup due to a small number of operations incapable of generating large amounts of interference (although large numbers of such operations can, in fact, lead to quantum speedup). Low-interference operations include sparse unitaries, Grover reflections, short-time and low-energy Hamiltonian evolutions, and the Haar wavelet transform. Circuits built from such operations can be classically simulated via a Monte Carlo technique making use of a convex combination of two Markov chains. Applications to query complexity, communication complexity, and the Wigner representation are discussed. © 2014 American Physical Society.


Singh P.V.,Carnegie Mellon University | Tan Y.,University of Washington | Mookerjee V.,University of Texas at Dallas
MIS Quarterly: Management Information Systems | Year: 2011

What determines the success of open source projects? In this study, we investigate the impact of network social capital on open source project success. We define network social capital as the benefits open source developers secure from their membership in developer collaboration networks. We focus on one specific type of success as measured by the rate of knowledge creation in an open source project. Specific hypotheses are developed and tested using a longitudinal panel of 2,378 projects hosted at SourceForge. We find that network social capital is not equally accessible to or appropriated by all projects. Our main results are as follows. First, projects with greater internal cohesion (that is, cohesion among the project members) are more successful. Second, external cohesion (that is, cohesion among the external contacts of a project) has an inverse U-shaped relationship with the project's success; moderate levels of external cohesion are best for a project's success rather than very low or very high levels. Third, the technological diversity of the external network of a project also has the greatest benefit when it is neither too low nor too high. Fourth, the number of direct and indirect external contacts positively affects a project's success such that the effect of the number of direct contacts is moderated by the number of indirect contacts. These results are robust to several control variables and alternate model specifications. Several theoretical and managerial implications are provided.


Sutner K.,Carnegie Mellon University
International Journal of General Systems | Year: 2012

We discuss attempts at the classification of cellular automata, in particular with a view towards decidability. We will see that a large variety of properties relating to the short-term evolution of configurations are decidable in principle, but questions relating to the long-term evolution are typically undecidable. Even in the decidable case, computational hardness poses a major obstacle for the automatic analysis of cellular automata. © 2012 Taylor & Francis.


Li C.,Tsinghua University | Kitani K.M.,Carnegie Mellon University
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition | Year: 2013

We address the task of pixel-level hand detection in the context of ego-centric cameras. Extracting hand regions in ego-centric videos is a critical step for understanding hand-object manipulation and analyzing hand-eye coordination. However, in contrast to traditional applications of hand detection, such as gesture interfaces or sign-language recognition, ego-centric videos present new challenges such as rapid changes in illuminations, significant camera motion and complex hand-object manipulations. To quantify the challenges and performance in this new domain, we present a fully labeled indoor/outdoor ego-centric hand detection benchmark dataset containing over 200 million labeled pixels, which contains hand images taken under various illumination conditions. Using both our dataset and a publicly available ego-centric indoors dataset, we give extensive analysis of detection performance using a wide range of local appearance features. Our analysis highlights the effectiveness of sparse features and the importance of modeling global illumination. We propose a modeling strategy based on our findings and show that our model outperforms several baseline approaches. © 2013 IEEE.


Garg R.,University of Texas at Austin | Telang R.,Carnegie Mellon University
MIS Quarterly: Management Information Systems | Year: 2013

With an abundance of products available online, many online retailers provide sales rankings to make it easier for consumers to find the best-selling products. Successfully implementing product rankings online was done a decade ago by Amazon, and more recently by Apple's App Store. However, neither market provides actual download data, a very useful statistic for both practitioners and researchers. In the past, researchers developed various strategies that allowed them to infer demand from rank data. Almost all of that work is based on an experiment that shifts sales or collaboration with a vendor to get actual sales data. In this research, we present an innovative method to use public data to infer the rank-demand relationship for the paid apps on Apple's iTunes App Store. We find that the top-ranked paid app for iPhone generates 150 times more downloads compared to the paid app ranked at 200. Similarly, the top paid app on iPad generates 120 times more downloads compared to the paid app ranked at 200. We conclude with a discussion on an extension of this framework to the Android platform, in-app purchases, and free apps.


Nagin D.S.,Carnegie Mellon University
Annals of Nutrition and Metabolism | Year: 2014

This article provides an overview of a group-based statistical methodology for analyzing developmental trajectories - the evolution of an outcome over age or time. Across all application domains, this group-based statistical method lends itself to the presentation of findings in the form of easily understood graphical and tabular data summaries. In so doing, the method provides statistical researchers with a tool for figuratively painting a statistical portrait of the predictors and consequences of distinct trajectories of development. Data summaries of this form have the great advantage of being accessible to nontechnical audiences and quickly comprehensible to audiences that are technically sophisticated. Examples of the application of the method are provided. A detailed account of the statistical underpinnings of the method and a full range of applications are provided by the author in a previous study. © 2014 S. Karger AG, Basel.


Abhishek V.,Carnegie Mellon University | Hosanagar K.,University of Pennsylvania
Operations Research | Year: 2013

We study optimal bidding strategies for advertisers in sponsored search auctions. In general, these auctions are run as variants of second-price auctions but have been shown to be incentive incompatible. Thus, advertisers have to be strategic about bidding. Uncertainty in the decision-making environment, budget constraints, and the presence of a large portfolio of keywords makes the bid optimization problem nontrivial. We present an analytical model to compute the optimal bids for keywords in an advertiser's portfolio. To validate our approach, we estimate the parameters of the model using data from an advertiser's sponsored search campaign and use the bids proposed by the model in a field experiment. The results of the field implementation show that the proposed bidding technique is very effective in practice. We extend our model to account for interactions between keywords, in the form of positive spillovers from generic keywords into branded keywords. The spillovers are estimated using a dynamic linear model framework and are used to jointly optimize the bids of the keywords using an approximate dynamic programming approach. Accounting for the interaction between keywords leads to an additional improvement in the campaign performance. © 2013 INFORMS.


Tsourakakis C.E.,Carnegie Mellon University
Knowledge and Information Systems | Year: 2011

Triangle counting is an important problem in graph mining. Two frequently used metrics in complex network analysis that require the count of triangles are the clustering coefficients and the transitivity ratio of the graph. Triangles have been used successfully in several real-world applications, such as detection of spamming activity, uncovering the hidden thematic structure of the web and link recommendation in online social networks. Furthermore, the count of triangles is a frequently used network statistic in exponential random graph models. However, counting the number of triangles in a graph is computationally expensive. In this paper, we propose the EigenTriangle and EigenTriangleLocal algorithms to estimate the number of triangles in a graph. The efficiency of our algorithms is based on the special spectral properties of real-world networks, which allow us to approximate accurately the number of triangles. We verify the efficacy of our method experimentally in almost 160 experiments using several Web Graphs, social, co-authorship, information, and Internet networks where we obtain significant speedups with respect to a straightforward triangle counting algorithm. Furthermore, we propose an algorithm based on Fast SVD which allows us to apply the core idea of the EigenTriangle algorithm on graphs which do not fit in the main memory. The main idea is a simple node-sampling process according to which node i is selected with probability d i/2m where d i is the degree of node i and m is the total number of edges in the graph. Our theoretical contributions also include a theorem that gives a closed formula for the number of triangles in Kronecker graphs, a model of networks which mimics several properties of real-world networks. © 2010 Springer-Verlag London Limited.


Peha J.M.,Carnegie Mellon University
Telecommunications Policy | Year: 2013

There has been considerable effort to let more wireless devices operate in white space spectrum, that is within frequency bands and geographic areas where no wireless devices are active. Making white space available is certainly useful, but there are other sharing opportunities as well, some of which have been obscured by dangerous misconceptions about the concept of unused spectrum. This paper discusses allowing more devices to operate safely in gray space spectrum, that is spectrum that is actively being used in that transmissions are underway - something many economic models assume is impossible. The paper focuses on primary-secondary sharing, so devices gaining access to spectrum operate on a secondary basis in a way that never causes harmful interference to primary systems. Examples of primary-secondary gray space sharing mechanisms are described in which devices are allowed to share spectrum with broadcasting, radar, and cellular systems. Quantitative analysis shows that it is technically possible to support significant communications among secondary devices in spectrum that is already heavily used by cellular or radar. However, gray space sharing generally causes primary and secondary systems to be more technically interdependent than white space sharing, so different policy and governance structures are needed. Secondary market rules can support gray space sharing in cases where there is a single primary spectrum user, such as a cellular carrier. In cases where technology is static, the regulator may be able to control access for secondary devices. However, in cases with multiple primary users and multiple secondary users of spectrum, as might be seen in bands with radar for example, a new kind of governance body will be needed to facilitate spectrum sharing. © 2012 Elsevier Ltd.


Jordan M.I.,University of California at Berkeley | Mitchell T.M.,Carnegie Mellon University
Science | Year: 2015

Machine learning addresses the question of how to build computers that improve automatically through experience. It is one of today's most rapidly growing technical fields, lying at the intersection of computer science and statistics, and at the core of artificial intelligence and data science. Recent progress in machine learning has been driven both by the development of new learning algorithms and theory and by the ongoing explosion in the availability of online data and low-cost computation. The adoption of data-intensive machine-learning methods can be found throughout science, technology and commerce, leading to more evidence-based decision-making across many walks of life, including health care, manufacturing, education, financial modeling, policing, and marketing.


Wesolowski A.,Carnegie Mellon University
PloS one | Year: 2012

The rapid adoption of mobile phone technologies in Africa is offering exciting opportunities for engaging with high-risk populations through mHealth programs, and the vast volumes of behavioral data being generated as people use their phones provide valuable data about human behavioral dynamics in these regions. Taking advantage of these opportunities requires an understanding of the penetration of mobile phones and phone usage patterns across the continent, but very little is known about the social and geographical heterogeneities in mobile phone ownership among African populations. Here, we analyze a survey of mobile phone ownership and usage across Kenya in 2009 and show that distinct regional, gender-related, and socioeconomic variations exist, with particularly low ownership among rural communities and poor people. We also examine patterns of phone sharing and highlight the contrasting relationships between ownership and sharing in different parts of the country. This heterogeneous penetration of mobile phones has important implications for the use of mobile technologies as a source of population data and as a public health tool in sub-Saharan Africa.


Matyjaszewski K.,Carnegie Mellon University | Tsarevsky N.V.,Southern Methodist University
Journal of the American Chemical Society | Year: 2014

This Perspective presents recent advances in macromolecular engineering enabled by ATRP. They include the fundamental mechanistic and synthetic features of ATRP with emphasis on various catalytic/initiation systems that use parts-per-million concentrations of Cu catalysts and can be run in environmentally friendly media, e.g., water. The roles of the major components of ATRP - monomers, initiators, catalysts, and various additives - are explained, and their reactivity and structure are correlated. The effects of media and external stimuli on polymerization rates and control are presented. Some examples of precisely controlled elements of macromolecular architecture, such as chain uniformity, composition, topology, and functionality, are discussed. Syntheses of polymers with complex architecture, various hybrids, and bioconjugates are illustrated. Examples of current and forthcoming applications of ATRP are covered. Future challenges and perspectives for macromolecular engineering by ATRP are discussed. © 2014 American Chemical Society.


Althoff M.,TU Munich | Krogh B.H.,Carnegie Mellon University
IEEE Transactions on Automatic Control | Year: 2014

This paper presents a numerical procedure for the reachability analysis of systems with nonlinear, semi-explicit, index-1 differential-algebraic equations. The procedure computes reachable sets for uncertain initial states and inputs in an overapproximative way, i.e. it is guaranteed that all possible trajectories of the system are enclosed. Thus, the result can be used for formal verification of system properties that can be specified in the state space as unsafe or goal regions. Due to the representation of reachable sets by zonotopes and the use of highly scalable operations on them, the presented approach scales favorably with the number of state variables. This makes it possible to solve problems of industry-relevant size, as demonstrated by a transient stability analysis of the IEEE 14-bus benchmark problem for power systems. © 1963-2012 IEEE.


In this paper, we optimize a process that integrates the use of glycerol to produce ethanol via fermentation within the simultaneous production of biodiesel and bioethanol from algae. The process consists of growing the algae, determining the optimal fraction of oil vs. starch, followed by oil extraction, starch liquefaction and saccharification, to sugars, oil transesterification, for which we consider two transesterification technologies (enzymes and alkali) and the fermentation of sugars and glycerol. The advantage of this process is that the dehydration technologies are common for the products of the glucose and glycerol fermentation. Simultaneous optimization and heat integration is performed using Duran and Grossmann's model. The fermentation of glycerol to ethanol increases the production of bioethanol by at least 50%. The energy and water consumptions are competitive with other processes that either sell the glycerol or use it to obtain methanol. However, the price for the biofuels is only competitive if glycerol cannot be sold to the market. © 2014 Elsevier Ltd.


Stern R.M.,Carnegie Mellon University | Morgan N.,UCB and ICSI
IEEE Signal Processing Magazine | Year: 2012

Many feature extraction methods that have been used for automatic speech recognition (ASR) have either been inspired by analogy to biological mechanisms, or at least have similar functional properties to biological or psychoacoustic properties for humans or other mammals. These methods have in many cases provided significant reductions in errors, particularly for degraded signals, and are currently experiencing a resurgence in community interest. Many of them have been quite successful, and others that are still in early stages of application still seem to hold great promise, given the existence proof of amazingly robust natural audio processing systems. © 2012 IEEE.


Armitage B.A.,Carnegie Mellon University
Current opinion in chemical biology | Year: 2011

Fluorescence microscopy and molecular tagging technologies have ushered in a new era in our understanding of protein localization and function in cells. This review summarizes recent efforts to extend some of these methods (and to create new ones) to imaging of RNA in live cells. Both fluorescent proteins and hybridization probes allow noncovalent labeling of specific RNA molecules with fluorescent dyes that allow detection and tracking in real time. Copyright © 2011 Elsevier Ltd. All rights reserved.


Yagan O.,Carnegie Mellon University
IEEE Transactions on Information Theory | Year: 2012

We investigate the secure connectivity of wireless sensor networks under the random key distribution scheme of Eschenauer and Gligor. Unlike recent work which was carried out under the assumption of full visibility, here we assume a (simplified) communication model where unreliable wireless links are represented as on/off channels. We present conditions on how to scale the model parameters so that the network: 1) has no secure node which is isolated and 2) is securely connected, both with high probability when the number of sensor nodes becomes large. The results are given in the form of full zero-one laws, and constitute the first complete analysis of the EG scheme under non-full visibility. Through simulations, these zero-one laws are shown to be valid also under a more realistic communication model (i.e., the disk model). The relations to the Gupta and Kumar's conjecture on the connectivity of geometric random graphs with randomly deleted edges are also discussed. © 1963-2012 IEEE.


Hong J.,Carnegie Mellon University
Communications of the ACM | Year: 2012

Looking past the systems people use, they target the people using the systems.


Bruchez M.P.,Carnegie Mellon University
Current opinion in chemical biology | Year: 2011

Thirteen years after the demonstration of quantum dots as biological imaging agents, and nine years after the initial commercial introduction of bioconjugated quantum dots, the brightness and photostability of the quantum dots has enabled a range of investigations using single molecule tracking. These materials are being routinely utilized by a number of groups to track the dynamics of single molecules in reconstituted biophysical systems and on living cells, and are especially powerful for investigations of single molecules over long timescales with short exposure times and high pointing accuracy. New approaches are emerging where the quantum dots are used as 'hard-sphere' probes for intracellular compartments. Innovations in quantum dot surface modification are poised to substantially expand the utility of these materials. Copyright © 2011 Elsevier Ltd. All rights reserved.


Redden J.P.,University of Minnesota | Galak J.,Carnegie Mellon University
Journal of Experimental Psychology: General | Year: 2013

The traditional view of satiation is that repeated consumption produces an unavoidable decline in liking according to the quantity and recency of consumption. We challenge this deterministic view by showing that satiation is instead partially constructed in the moment based on contextual cues. More specifically, while satiation is a function of the actual amount consumed, it also depends on the subjective sense of how much one has recently consumed. We demonstrate the influence of this subjective sense of satiation and show that it is driven by metacognitive cues such as the ease of retrieval of past experiences (Experiments 1 and 2) and can also be directly manipulated by providing a normative standard for consumption quantity (Experiment 3). Our research demonstrates that satiety is not driven solely by the amount and timing of past consumption, thereby establishing the role of higher order metacognitive inferences in satiation and providing insight into how they underlie the construction of satiation. © 2012 American Psychological Association.


Ahmetovic E.,University of Tuzla | Grossmann I.E.,Carnegie Mellon University
AIChE Journal | Year: 2011

We propose a general superstructure and a model for the global optimization for integrated process water networks. The superstructure consists of multiple sources of water, water-using processes, wastewater treatment, and pre-treatment operations. Unique features are that all feasible interconnections are considered between them and multiple sources of water can be used. The proposed model is formulated as a nonlinear programing (NLP) and as a mixed integer nonlinear programing (MINLP) problem for the case when 0-1 variables are included for the cost of piping and to establish optimal trade-offs between cost and network complexity. To effectively solve the NLP and MINLP models to global optimality we propose tight bounds on the variables, which are expressed as general equations. We also incorporate the cut proposed by Karuppiah and Grossmann to significantly improve the strength of the lower bound for the global optimum. The proposed model is tested on several examples. © 2010 American Institute of Chemical Engineers (AIChE).


Kuwahara H.,Carnegie Mellon University | Soyer O.S.,University of Exeter
Molecular Systems Biology | Year: 2012

Noisy bistable dynamics in gene regulation can underlie stochastic switching and is demonstrated to be beneficial under fluctuating environments. It is not known, however, if fluctuating selection alone can result in bistable dynamics. Using a stochastic model of simple feedback networks, we apply fluctuating selection on gene expression and run in silico evolutionary simulations. We find that independent of the specific nature of the environment-fitness relationship, the main outcome of fluctuating selection is the evolution of increased evolvability in the network; system parameters evolve toward a nonlinear regime where phenotypic diversity is increased and small changes in genotype cause large changes in expression level. In the presence of noise, the evolution of increased nonlinearity results in the emergence and maintenance of bistability. Our results provide the first direct evidence that bistability and stochastic switching in a gene regulatory network can emerge as a mechanism to cope with fluctuating environments. They strongly suggest that such emergence occurs as a byproduct of evolution of evolvability and exploitation of noise by evolution. © 2012 EMBO and Macmillan Publishers Limited All rights reserved.


De Graef M.,Carnegie Mellon University
Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science | Year: 2010

The direct visualization of crystallographic and magnetic point group symmetry by means of computer graphics simplifies the teaching of point group symmetry at the undergraduate and graduate levels. Still images and animated movies for all 32 crystallographic point groups, and for the 122 magnetic point groups, are presented. For each point group, the action of the symmetry operators on scalar and pseudo-scalar objects, as well as on polar and axial vectors, is represented as a three-dimensional rendered image. All images and movies are made available as supplementary educational material via a dedicated web site. © 2010 The Minerals, Metals & Materials Society and ASM International.


Hahamy A.,Weizmann Institute of Science | Behrmann M.,Carnegie Mellon University | Malach R.,Weizmann Institute of Science
Nature Neuroscience | Year: 2015

Autism spectrum disorder (ASD) has been associated with a reduction in resting state functional connectivity, though this assertion has recently been challenged by reports of increased connectivity in ASD. To address these contradictory findings, we examined both inter-and intrahemispheric functional connectivity in several resting state data sets acquired from adults with high-functioning ASD and matched control participants. Our results reveal areas of both increased and decreased connectivity in multiple ASD groups as compared to control groups. We propose that this heterogeneity stems from a previously unrecognized ASD characteristic: idiosyncratic distortions of the functional connectivity pattern relative to the typical, canonical template. The magnitude of an individual's pattern distortion in homotopic interhemispheric connectivity correlated significantly with behavioral symptoms of ASD. We propose that individualized alterations in functional connectivity organization are a core characteristic of high-functioning ASD, and that this may account for previous discrepant findings. © 2015 Nature America, Inc. All rights reserved.


Thiessen E.D.,Carnegie Mellon University
Child Development | Year: 2011

All theories of language development suggest that learning is constrained. However, theories differ on whether these constraints arise from language-specific processes or have domain-general origins such as the characteristics of human perception and information processing. The current experiments explored constraints on statistical learning of patterns, such as the phonotactic patterns of an infants' native language. Infants in these experiments were presented with a visual analog of a phonotactic learning task used by J. R. Saffran and E. D. Thiessen (2003). Saffran and Thiessen found that infants' phonotactic learning was constrained such that some patterns were learned more easily than other patterns. The current results indicate that infants' learning of visual patterns shows the same constraints as infants' learning of phonotactic patterns. This is consistent with theories suggesting that constraints arise from domain-general sources and, as such, should operate over many kinds of stimuli in addition to linguistic stimuli. © 2011 The Author. Child Development © 2011 Society for Research in Child Development, Inc.


Fischhoff B.,Carnegie Mellon University
Science | Year: 2015

Formal analyses can be valuable aids to decision-making if their limits are understood. Those limits arise from the two forms of subjectivity found in all analyses: ethical judgments, made when setting the terms of an analysis, and scientific judgments, made when conducting it. As formal analysis has assumed a larger role in policy decisions, awareness of those judgments has grown, as have methods for making them. The present review traces these developments, using examples that illustrate the issues that arise when designing, executing, and interpreting analyses. It concludes with lessons learned from the science and practice of analysis. One common thread in these lessons is the importance of collaborative processes, whereby analysts and decision-makers educate one another about their respective needs and capabilities.


Vaidya V.,Carnegie Mellon University
Physical Review D - Particles, Fields, Gravitation and Cosmology | Year: 2015

We utilize generalized unitarity and recursion relations combined with effective field theory techniques to compute spin-dependent interaction terms for an inspiralling binary system in the post-Newtonian (PN) approximation. Using these methods offers great computational advantage over traditional techniques involving Feynman diagrams, especially at higher orders in the PN expansion. As a specific example, we reproduce the spin-orbit (up to 2.5PN order) and the leading-order S2 (2PN) Hamiltonian for a binary system with one of the massive objects having nonzero spin using the S-matrix elements of elementary particles. For the same system, we also obtain the S3 (3.5PN) spin Hamiltonian for an arbitrary massive object, which was until now known only for a black hole. Furthermore, we derive the missing S4 Hamiltonian at leading order (4PN), again for an arbitrary massive object and establish that the minimal coupling of an elementary particle to gravity automatically captures the physics of a spinning black hole. Finally, the Kerr metric is obtained as a series in GN by comparing the action of a test particle in the vicinity of a spinning black hole to the derived potential. © 2015 American Physical Society.


Pressman S.D.,University of Kansas | Cohen S.,Carnegie Mellon University
Health Psychology | Year: 2012

Objective: This study examined whether specific types of positive and negative emotional words used in the autobiographies of well-known deceased psychologists were associated with longevity. Methods: For each of the 88 psychologists, the percent of emotional words used in writing was calculated and categorized by valence (positive or negative) and arousal (activated [e.g., lively, anxious] or not activated [e.g., calm, drowsy]) based on existing emotion scales and models of emotion categorization. Results: After controlling for sex, year of publication, health (based on disclosed illness in autobiography), native language, and year of birth, the use of more activated positive emotional words (e.g., lively, vigorous, attentive, humorous) was associated with increased longevity. Negative terms (e.g., angry, afraid, drowsy, sluggish) and unactivated positive terms (e.g., peaceful, calm) were not related to longevity. The association of activated positive emotions with longevity was also independent of words indicative of social integration, optimism, and the other affect/activation categories. Conclusions: Results indicate that in writing, not every type of emotion correlates with longevity and that there may be value to considering different categories beyond emotional valence in health relevant outcomes. © 2011 American Psychological Association.


Natarajan A.,Carnegie Mellon University
Physical Review D - Particles, Fields, Gravitation and Cosmology | Year: 2012

We use cosmic microwave background data from the WMAP, SPT, BICEP, and QUaD experiments to obtain constraints on the dark matter particle mass m χ, and show that the combined data requires m χ> 7.6GeV at the 95% confidence level for the χχ→bb̄ channel assuming s-wave annihilation and a thermal cross section σ av= 3×10 -26cm3/s. We examine whether the bound on m χ is sensitive to σ 8 measurements made by galaxy cluster observations. The large uncertainty in σ 8 and the degeneracy with Ω m allow only small improvements in the dark matter mass bound. Increasing the number of effective neutrinolike degrees of freedom to N eff=3.85 improves the mass bound to m χ>8.6GeV at 95% confidence, for the χχ→bb̄ channel. We also study models in which dark matter halos at z<60 reionize the Universe. We compute the Ostriker-Vishniac power resulting from partial reionization at intermediate redshifts 10


Sokolov A.N.,Stanford University | Tee B.C.-K.,Stanford University | Bettinger C.J.,Carnegie Mellon University | Tok J.B.-H.,Stanford University | Bao Z.,Stanford University
Accounts of Chemical Research | Year: 2012

Figure Persented: Skin is the body's largest organ and is responsible for the transduction of a vast amount of information. This conformable material simultaneously collects signals from external stimuli that translate into information such as pressure, pain, and temperature. The development of an electronic material, inspired by the complexity of this organ is a tremendous, unrealized engineering challenge. However, the advent of carbon-based electronics may offer a potential solution to this long-standing problem.In this Account, we describe the use of an organic field-effect transistor (OFET) architecture to transduce mechanical and chemical stimuli into electrical signals. In developing this mimic of human skin, we thought of the sensory elements of the OFET as analogous to the various layers and constituents of skin. In this fashion, each layer of the OFET can be optimized to carry out a specific recognition function. The separation of multimodal sensing among the components of the OFET may be considered a "divide and conquer" approach, where the electronic skin (e-skin) can take advantage of the optimized chemistry and materials properties of each layer.This design of a novel microstructured gate dielectric has led to unprecedented sensitivity for tactile pressure events. Typically, pressure-sensitive components within electronic configurations have suffered from a lack of sensitivity or long mechanical relaxation times often associated with elastomeric materials. Within our method, these components are directly compatible with OFETs and have achieved the highest reported sensitivity to date. Moreover, the tactile sensors operate on a time scale comparable with human skin, making them ideal candidates for integration as synthetic skin devices. The methodology is compatible with large-scale fabrication and employs simple, commercially available elastomers.The design of materials within the semiconductor layer has led to the incorporation of selectivity and sensitivity within gas-sensing devices and has enabled stable sensor operation within aqueous media. Furthermore, careful tuning of the chemical composition of the dielectric layer has provided a means to operate the sensor in real time within an aqueous environment and without the need for encapsulation layers.The integration of such devices as electronic mimics of skin will require the incorporation of biocompatible or biodegradable components. Toward this goal, OFETs may be fabricated with >99% biodegradable components by weight, and the devices are robust and stable, even in aqueous environments. Collectively, progress to date suggests that OFETs may be integrated within a single substrate to function as an electronic mimic of human skin, which could enable a large range of sensing-related applications from novel prosthetics to robotic surgery. © 2011 American Chemical Society.


Ettensohn C.A.,Carnegie Mellon University
Genesis | Year: 2013

A central challenge of developmental and evolutionary biology is to explain how anatomy is encoded in the genome. Anatomy emerges progressively during embryonic development, as a consequence of morphogenetic processes. The specialized properties of embryonic cells and tissues that drive morphogenesis, like other specialized properties of cells, arise as a consequence of differential gene expression. Recently, gene regulatory networks (GRNs) have proven to be powerful conceptual and experimental tools for analyzing the genetic control and evolution of developmental processes. A major current goal is to link these transcriptional networks directly to morphogenetic processes. This review highlights three experimental models (sea urchin skeletogenesis, ascidian notochord morphogenesis, and the formation of somatic muscles in Drosophila) that are currently being used to analyze the genetic control of anatomy by integrating information of several important kinds: (1) morphogenetic mechanisms at the molecular, cellular and tissue levels that are responsible for shaping a specific anatomical feature, (2) the underlying GRN circuitry deployed in the relevant cells, and (3) modifications to gene regulatory circuitry that have accompanied evolutionary changes in the anatomical feature. © 2013 Wiley Periodicals, Inc.


Guruswami V.,Carnegie Mellon University | Xing C.,Nanyang Technological University
Proceedings of the Annual ACM Symposium on Theory of Computing | Year: 2013

We consider Reed-Solomon (RS) codes whose evaluation points belong to a subfield, and give a linear-algebraic list decoding algorithm that can correct a fraction of errors approaching the code distance, while pinning down the candidate messages to a well-structured affine space of dimension a constant factor smaller than the code dimension. By pre-coding the message polynomials into a subspace-evasive set, we get a Monte Carlo construction of a subcode of Reed-Solomon codes that can be list decoded from a fraction (1 - R - ε) of errors in polynomial time (for any fixed ε > 0) with a list size of O(1=ε). Our methods extend to algebraic-geometric (AG) codes, leading to a similar claim over constant-sized alphabets. This matches parameters of recent results based on folded variants of RS and AG codes. Further, the underlying algebraic idea also extends nicely to Gabidulin's construction of rank-metric codes based on linearized polynomials. This gives the first construction of positive rate rank-metric codes list decodable beyond half the distance, and in fact gives codes of rate R list decodable up to the optimal (1 - R - ε) fraction of rank errors. We introduce a new notion called subspace designs as another way to pre-code messages and prune the subspace of candidate solutions. Using these, we also get a deterministic construction of a polynomial time list decodable subcode of RS codes. By using a cascade of several subspace designs, we extend our approach to AG codes, which gives the first deterministic construction of an algebraic code family of rate R with efficient list decoding from 1-R-ε fraction of errors over an alphabet of constant size (that depends only on ε). The list size bound is almost a constant (governed by log (block length)), and the code can be constructed in quasi-polynomial time. Copyright 2013 ACM.


Bruine De Bruin W.,University of Leeds | Bruine De Bruin W.,Carnegie Mellon University | Bostrom A.,University of Washington
Proceedings of the National Academy of Sciences of the United States of America | Year: 2013

As members of a democratic society, individuals face complex decisions about whether to support climate change mitigation, vaccinations, genetically modified food, nanotechnology, geoengineering, and so on. To inform people's decisions and public debate, scientific experts at government agencies, nongovernmental organizations, and other organizations aim to provide understandable and scientifically accurate communication materials. Such communications aim to improve people's understanding of the decision-relevant issues, and if needed, promote behavior change. Unfortunately, existing communications sometimes fail when scientific experts lack information about what people need to know to make more informed decisions or what wording people use to describe relevant concepts. We provide an introduction for scientific experts about how to use mental models research with intended audience members to inform their communication efforts. Specifically, we describe how to conduct interviews to characterize people's decision-relevant beliefs or mental models of the topic under consideration, identify gaps and misconceptions in their knowledge, and reveal their preferred wording. We also describe methods for designing follow-up surveys with larger samples to examine the prevalence of beliefs as well as the relationships of beliefs with behaviors. Finally, we discuss how findings from these interviews and surveys can be used to design communications that effectively address gaps and misconceptions in people's mental models in wording that they understand. We present applications to different scientific domains, showing that this approach leads to communications that improve recipients' understanding and ability to make informed decisions.


Noise-resistant and nuanced, COGBASE makes 10 million pieces of commonsense data and a host of novel reasoning algorithms available via a family of semantically-driven prior probability distributions.Machine learning, Big Data, natural language understanding/processing, and social AI can draw on COGBASE to determine lexical semantics, infer goals and interests, simulate emotion and affect, calculate document gists and topic models, and link commonsense knowledge to domain models and social, spatial, cultural, and psychological data.COGBASE is especially ideal for social Big Data, which tends to involve highly implicit contexts, cognitive artifacts, difficult-to-parse texts, and deep domain knowledge dependencies. © 2014 Elsevier Ltd.


Cimpian A.,University of Illinois at Urbana - Champaign | Erickson L.C.,Carnegie Mellon University
Cognitive Psychology | Year: 2012

What are the representations and learning mechanisms that underlie conceptual development? The present research provides evidence in favor of the claim that this process is guided by an early-emerging predisposition to think and learn about abstract kinds. Specifically, three studies (N=192) demonstrated that 4- to 7-year-old children have better recall for novel information about kinds (e.g., that dogs catch a bug called " fep" ) than for similar information about individuals (e.g., that a particular dog catches a bug called " fep" ). By showing that children are particularly likely to retain information about kinds, this work not only provides a first empirical demonstration of a phenomenon that may be key to conceptual development but also makes it apparent that young children's thinking is suffused with abstractions rather than being perceptually-based and concrete. © 2011 Elsevier Inc.


Meyer C.A.,Carnegie Mellon University | Swanson E.S.,University of Pittsburgh
Progress in Particle and Nuclear Physics | Year: 2015

A review of the theoretical and experimental status of hybrid hadrons is presented. The states π1(1400), π1(1600), and π1(2015) are thoroughly reviewed, along with experimental results from GAMS, VES, Obelix, COMPASS, KEK, CLEO, Crystal Barrel, CLAS, and BNL. Theoretical lattice results on the gluelump spectrum, adiabatic potentials, heavy and light hybrids, and transition matrix elements are discussed. These are compared with bag, string, flux tube, and constituent gluon models. Strong and electromagnetic decay models are described and compared to lattice gauge theory results. We conclude that while good evidence for the existence of a light isovector exotic meson exists, its confirmation as a hybrid meson awaits discovery of its iso-partners. We also conclude that lattice gauge theory rules out a number of hybrid models and provides a reference to judge the success of others. © 2015 Elsevier B.V.All rights reserved.


McNew J.A.,Rice University | Sondermann H.,Cornell University | Lee T.,Carnegie Mellon University | Stern M.,Rice University | Brandizzi F.,Michigan State University
Annual Review of Cell and Developmental Biology | Year: 2013

Shape changes and topological remodeling of membranes are essential for the identity of organelles and membrane trafficking. Although all cellular membranes have common features, membranes of different organelles create unique environments that support specialized biological functions. The endoplasmic reticulum (ER) is a prime example of this specialization, as its lipid bilayer forms an interconnected system of cisternae, vesicles, and tubules, providing a highly compartmentalized structure for a multitude of biochemical processes. A variety of peripheral and integral membrane proteins that facilitate membrane curvature generation, fission, and/or fusion have been identified over the past two decades. Among these, the dynamin-related proteins (DRPs) have emerged as key players. Here, we review recent advances in our functional and molecular understanding of fusion DRPs, exemplified by atlastin, an ER-resident DRP that controls ER structure, function, and signaling. © 2013 by Annual Reviews. All rights reserved.


Ross A.,Carnegie Mellon University
Physical Review D - Particles, Fields, Gravitation and Cosmology | Year: 2012

Sources of long wavelength radiation are naturally described by an effective field theory (EFT) which takes the form of a multipole expansion. Its action is given by a derivative expansion where higher order terms are suppressed by powers of the ratio of the size of the source over the wavelength. In order to determine the Wilson coefficients of the EFT, i.e. the multipole moments, one needs the mapping between a linear source term action and the multipole expansion form of the action of the EFT. In this paper we perform the multipole expansion to all orders by Taylor expanding the field in the source term and then decomposing the action into symmetric trace-free tensors which form irreducible representations of the rotation group. We work at the level of the action, and we obtain the action to all orders in the multipole expansion and the exact expressions for the multipole moments for a scalar field, electromagnetism, and linearized gravity. Our results for the latter two cases are manifestly gauge invariant. We also give expressions for the energy flux and the (gauge dependent) radiation field to all orders in the multipole expansion. The results for linearized gravity are a component of the EFT framework NRGR and will greatly simplify future calculations of gravitational wave observables in the radiation sector of NRGR. © 2012 American Physical Society.


Matyjaszewski K.,Carnegie Mellon University
Macromolecules | Year: 2012

Current status and future perspectives in atom transfer radical polymerization (ATRP) are presented. Special emphasis is placed on mechanistic understanding of ATRP, recent synthetic and process development, and new controlled polymer architectures enabled by ATRP. New hybrid materials based on organic/inorganic systems and natural/synthetic polymers are presented. Some current and forthcoming applications are described. © 2012 American Chemical Society.


Chen C.-Y.,Carnegie Mellon University | Dev P.S.B.,University of Maryland University College
Physical Review D - Particles, Fields, Gravitation and Cosmology | Year: 2012

We discuss the possibility of observing multi-lepton signals at the Large Hadron Collider (LHC) from the production and decay of heavy standard model (SM) singlet neutrinos added in extensions of SM to explain the observed light neutrino masses by seesaw mechanism. In particular, we analyze two "smoking gun" signals depending on the Dirac or Majorana nature of the heavy neutrino: (i) for Majorana case, the same-sign di-lepton signal which can be used as a probe of lepton-number violation, and (ii) for Dirac case, the tri-lepton signal which conserves lepton number but may violate lepton flavor. Within a minimal Left-Right symmetric framework in which these additional neutrino states arise naturally, we find that in both cases, the signals can be identified with virtually no background beyond a TeV, and the heavy gauge boson W R can be discovered in this process. This analysis also provides a direct way to probe the nature of seesaw physics involving the SM singlets at TeV-scale, and in particular, to distinguish type-I seesaw with purely Majorana heavy neutrinos from inverse seesaw with pseudo-Dirac counterparts. © 2012 American Physical Society.


Walker R.S.,University of Missouri | Bailey D.H.,Carnegie Mellon University
Evolution and Human Behavior | Year: 2013

Violence was likely often a strong selective pressure in many traditional lowland South American societies. A compilation of 11 anthropological studies reporting cause of death shows that violence led to about 30% of adult deaths, of which about 70% were males. Here violent deaths are further itemized at the level of ethnographically-reported death events (particular duels, homicides, and raids) to provide more detailed insight into the causes and consequences of within- and between-group violence. Data for 238 death events (totaling 1145 deaths) from 44 lowland South American societies show that attacks are more deadly when treachery is used, when avenging a previous killing, and on external warfare raids between ethnolinguistic groups. That revenge raids kill more people on average than the original grievance, at least when conflicts are between ethnolinguistic groups, indicates a tendency towards increasingly vicious cycles of revenge killings. Motives of killings as noted in ethnographic sources, in order of importance, reportedly include revenge for previous killings and other wrong-doings like sorcery, jealousy over women, gain of captive women and children, fear or deterrence of impending attack, and occasionally the theft of material goods. Results may have implications for understanding the potential for multi-level selection by delineating the force of competition at varying scales of analysis within and between lowland South American societies. © 2013 Elsevier Inc.


Cataldo M.,Robert Bosch GmbH | Herbsleb J.D.,Carnegie Mellon University
IEEE Transactions on Software Engineering | Year: 2013

The success of software development projects depends on carefully coordinating the effort of many individuals across the multiple stages of the development process. In software engineering, modularization is the traditional technique intended to reduce the interdependencies among modules that constitute a system. Reducing technical dependencies, the theory argues, results in a reduction of work dependencies between teams developing interdependent modules. Although that research stream has been quite influential, it considers a static view of the problem of coordination in engineering activities. Building on a dynamic view of coordination, we studied the relationship between socio-technical congruence and software quality and development productivity. In order to investigate the generality of our findings, our analyses were performed on two large-scale projects from two companies with distinct characteristics in terms of product and process maturity. Our results revealed that the gaps between coordination requirements and the actual coordination activities carried out by the developers significantly increased software failures. Our analyses also showed that higher levels of congruence are associated with improved development productivity. Finally, our results showed the congruence between dependencies and coordinative actions is critical both in mature development settings as well as in novel and dynamic development contexts. © 1976-2012 IEEE.


Bruine De Bruin W.,Carnegie Mellon University | Carman K.G.,University of Tilburg
Medical Decision Making | Year: 2012

Objectives. Risk perceptions are central to good health decisions. People can judge valid probabilities but use 50% disproportionately. The authors hypothesized that 50% is more likely than other responses to reflect not knowing the probability, especially among individuals with low education and numeracy, and evaluated the usefulness of eliciting "don't know" explanations. Methods. Respondents (n = 1020) judged probabilities for living or dying in the next 10 years, indicating whether they gave a good estimate or did not know the chances. They completed demographics, medical history, and numeracy questions. Results. Overall, 50% was more likely than other probabilities to be explained as "don't know" (v. "a good estimate"). Correlations of using 50% with low education and numeracy were mediated by expressing "don't know." Judged probabilities for survival and mortality explained as "don't know" had lower correlations with age, diseases, and specialist visits. Conclusions. When judging risks, 50% may reflect not knowing the probability, especially among individuals with low numeracy and education. Probabilities expressed as "don't know" are less valid. Eliciting uncertainty could benefit theoretical models and educational efforts.


Feinberg A.W.,Carnegie Mellon University
Annual Review of Biomedical Engineering | Year: 2015

In nature, nanometer-scale molecular motors are used to generate force within cells for diverse processes from transcription and transport to muscle contraction. This adaptability and scalability across wide temporal, spatial, and force regimes have spurred the development of biological soft robotic systems that seek to mimic and extend these capabilities. This review describes how molecular motors are hierarchically organized into larger-scale structures in order to provide a basic understanding of how these systems work in nature and the complexity and functionality we hope to replicate in biological soft robotics. These span the subcellular scale to macroscale, and this article focuses on the integration of biological components with synthetic materials, coupled with bioinspired robotic design. Key examples include nanoscale molecular motor-powered actuators, microscale bacteria-controlled devices, and macroscale muscle-powered robots that grasp, walk, and swim. Finally, the current challenges and future opportunities in the field are addressed. © 2015 by Annual Reviews. All rights reserved.


Sandholm T.,Carnegie Mellon University
AI Magazine | Year: 2010

Incomplete-information games are games where the agents do not entirely know the state of the game at all times. Poker has emerged as a standard benchmark in this space for a number of reasons because it exhibits the richness of reasoning about a probabilistic future, how to interpret others' actions as signals, and information hiding through careful action selection, and the game can be scaled to the desired complexity. Most competitive poker-playing programs are generated using an abstraction algorithm followed by using a custom equilibrium finding algorithm to solve the abstracted game. Some games are so large that even after applying the kind of lossless abstraction described above, the resulting LP would be too large to solve. To address this problem, such games can be abstracted more aggressively, but this incurs loss in solution quality. Nash equilibrium provides a safety guarantee in two-player zero-sum games.


London A.J.,Carnegie Mellon University
Journal of Law, Medicine and Ethics | Year: 2012

This paper offers a non-paternalistic justification for prospective research review as providing a credible social assurance that the institutions of scientific advancement respect and affirm the moral equality of all community members and as creating a "market" in which stakeholders working to advance diverse ends also advance the common good. © 2012 American Society of Law, Medicine & Ethics, Inc.


Gross M.D.,Carnegie Mellon University
Personal and Ubiquitous Computing | Year: 2011

As we rapidly approach the day of transitive materials, made of individual elements that sense and actuate and can be programmed and reprogrammed, it is time to think about how to design things using these new materials. Our roBlocks construction kit toy teaching children about emergent behavior in complex systems serves as an example for investigating the challenges of designing things made of transitive materials. The roBlocks kit comprises heterogeneous modular robotics components that exhibit modularity, one-to-one mapping between form and behavior, and non-hierarchical control; and these features make it appropriate for experimenting with emergent behavior. However, as the numbers of robotic components scales to the orders of magnitude needed to consider them as material these same features also make it difficult to apply traditional methods to design constructions with desired behaviors. To understand this design space we built, the Erstwhile Agent that uses an evolutionary approach to automatically synthesize roBlocks constructions to meet specified desiderata. © 2010 Springer-Verlag London Limited.


Beladi H.,Deakin University | Rohrer G.S.,Carnegie Mellon University
Acta Materialia | Year: 2013

The relative grain boundary area and energy distributions of a ferritic steel were characterized as a function of lattice misorientation and boundary plane orientation using focused ion beam serial sectioning combined with electron backscatter diffraction. The grain boundary energy and population depended on both the grain boundary plane orientation and lattice misorientation. When misorientation was ignored grain boundary planes with the (1 1 1) orientation had the minimum energy and the largest relative areas. The most commonly observed boundaries were {1 1 2} symmetric tilt boundaries with the Σ3 misorientation; this boundary also had a low energy. On average there was a strong inverse correlation between the relative areas of different types of grain boundaries and the relative grain boundary energies. © 2012 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.


Swendsen R.H.,Carnegie Mellon University
American Journal of Physics | Year: 2011

Discussions of the foundations of statistical mechanics, how they lead to thermodynamics, and the appropriate definition of entropy have occasioned many disagreements. I believe that some or all of these disagreements arise from differing, but unstated assumptions, which can make opposing opinions difficult to reconcile. To make these assumptions explicit, I discuss the principles that have guided my own thinking about the foundations of statistical mechanics, the microscopic origins of thermodynamics, and the definition of entropy. The purpose of this paper will be fulfilled if it paves the way to a final consensus, whether or not that consensus agrees with my point of view. © 2011 American Association of Physics Teachers.


Griffiths R.B.,Carnegie Mellon University
American Journal of Physics | Year: 2011

Maudlin has claimed that no local theory can reproduce the predictions of standard quantum mechanics that violate Bell's inequality for Bohm's version (two spin-half particles in a singlet state) of the Einstein-Podolsky-Rosen problem. It is argued that, on the contrary, standard quantum mechanics itself is a counterexample to Maudlin's claim, because it is local in the appropriate sense (measurements at one place do not influence what occurs elsewhere there) when formulated using consistent principles in place of the inconsistent appeals to "measurement" found in current textbooks. This argument sheds light on the claim of Blaylock that counterfactual definiteness is an essential ingredient in derivations of Bell's inequality. © 2011 American Association of Physics Teachers.


Bahar I.,University of Pittsburgh | Bahar I.,Carnegie Mellon University | Lezon T.R.,University of Pittsburgh | Bakan A.,University of Pittsburgh | Shrivastava I.H.,University of Pittsburgh
Chemical Reviews | Year: 2010

A normal mode analysis (NMA) of biomolecular structures and functional mechanisms of membrane proteins was studied. Normal mode analysis provides information on the equilibrium modes accessible to a system, assuming that the system is stabilized by harmonic potentials. A striking feature of NMA and other PCAs of biomolecular structures is the observed robustness of the global modes to details in atomic coordinates or specific interatomic interactions. A major reason behind the broadening recognition of NMA as a tool for exploring functional motions of proteins is the observation that global modes elucidated by NMA bear functional significance. Membrane proteins are classified into two broad groups, integral membrane proteins (IMPs) and peripheral membrane proteins. The biological function of many membrane proteins involves a transient change in structure, with the associated processes usually spanning a broad range of time scales. The function of membrane proteins involves many other specific and subtle interactions that cannot be studied by CG models and NMA.


Gurtin M.E.,Carnegie Mellon University | Ohno N.,Nagoya University
Journal of the Mechanics and Physics of Solids | Year: 2011

This paper develops a gradient theory of single-crystal plasticity based on a system of microscopic force balances, one balance for each slip system, derived from the principle of virtual power, and a mechanical version of the second law that includes, via the microscopic forces, work performed during plastic flow. When combined with thermodynamically consistent constitutive relations the microscopic force balances become nonlocal flow rules for the individual slip systems in the form of partial differential equations requiring boundary conditions. Central ingredients in the theory are densities of (geometrically necessary) edge and screw dislocations, densities that describe the accumulation of dislocations, and densities that characterize forest hardening. The form of the forest densities is based on an explicit kinematical expression for the normal Burgers vector on a slip plane. © 2010 2010 Elsevier Ltd. All rights reserved.


Ramanathan A.,Carnegie Mellon University | Ramanathan A.,Oak Ridge National Laboratory | Agarwal P.K.,Oak Ridge National Laboratory
PLoS Biology | Year: 2011

Proteins are intrinsically flexible molecules. The role of internal motions in a protein's designated function is widely debated. The role of protein structure in enzyme catalysis is well established, and conservation of structural features provides vital clues to their role in function. Recently, it has been proposed that the protein function may involve multiple conformations: the observed deviations are not random thermodynamic fluctuations; rather, flexibility may be closely linked to protein function, including enzyme catalysis. We hypothesize that the argument of conservation of important structural features can also be extended to identification of protein flexibility in interconnection with enzyme function. Three classes of enzymes (prolyl-peptidyl isomerase, oxidoreductase, and nuclease) that catalyze diverse chemical reactions have been examined using detailed computational modeling. For each class, the identification and characterization of the internal protein motions coupled to the chemical step in enzyme mechanisms in multiple species show identical enzyme conformational fluctuations. In addition to the active-site residues, motions of protein surface loop regions (&10 Å away) are observed to be identical across species, and networks of conserved interactions/residues connect these highly flexible surface regions to the active-site residues that make direct contact with substrates. More interestingly, examination of reaction-coupled motions in non-homologous enzyme systems (with no structural or sequence similarity) that catalyze the same biochemical reaction shows motions that induce remarkably similar changes in the enzyme-substrate interactions during catalysis. The results indicate that the reaction-coupled flexibility is a conserved aspect of the enzyme molecular architecture. Protein motions in distal areas of homologous and non-homologous enzyme systems mediate similar changes in the active-site enzyme-substrate interactions, thereby impacting the mechanism of catalyzed chemistry. These results have implications for understanding the mechanism of allostery, and for protein engineering and drug design. © 2011 Ramanathan, Agarwal.


Coles P.J.,Carnegie Mellon University
Physical Review A - Atomic, Molecular, and Optical Physics | Year: 2012

Macroscopic behavior such as the lack of interference patterns has been attributed to "decoherence," a word with several possible definitions such as (1) the loss of off-diagonal density matrix elements, (2) the flow of information to the environment, (3) the loss of complementary information, and (4) the loss of the ability to create entanglement in a measurement. In this article, we attempt to unify these distinct definitions by providing general quantitative connections between them, valid for all finite-dimensional quantum systems or quantum processes. The most important application of our results is to the understanding of quantum discord, a measure of the nonclassicality of the correlations between quantum systems. We show that some popular measures of discord measure the information missing from the purifying system and hence quantify security, which can be stated operationally in terms of distillable secure bits. The results also give some strategies for constructing discord measures. © 2012 American Physical Society.


Minden J.S.,Carnegie Mellon University
Methods in Molecular Biology | Year: 2012

Two-dimensional difference gel electrophoresis (2D DIGE) is a modified form of 2D electrophoresis (2DE) that allows one to compare two or three protein samples simultaneously on the same gel. The proteins in each sample are covalently tagged with different color fluorescent dyes that are designed to have no effect on the relative migration of proteins during electrophoresis. Proteins that are common to the samples appear as "spots" with a fixed ratio of fluorescent signals, whereas proteins that differ between the samples have different fluorescence ratios. With the appropriate imaging system, difference gel electrophoresis (DIGE) is capable of reliably detecting as little as 0.2 fmol of protein, and protein differences down to ±15%, over a ∼20,000-fold protein concentration range. DIGE combined with digital image analysis therefore greatly improves the statistical assessment of proteome variation. Here we describe a protocol for conducting DIGE experiments, which takes 2-3 days to complete. © 2012 Springer Science+Business Media, LLC.


Gaynor M.,Federal Trade Commission | Gaynor M.,Carnegie Mellon University
Health Affairs | Year: 2014

US health care is in ferment. Private entities are merging, aligning, and coordinating in a wide array of configurations. At the same time, there is a great deal of policy change. This includes the federal government's Affordable Care Act, as well as actions by Medicare, state legislatures, and state agencies. The health system is built upon markets, which determine how (and how well) goods and services are delivered to consumers, so it is critical that these markets work as well as possible. As the primary federal antitrust enforcement agencies, the Federal Trade Commission and the Department of Justice are charged with ensuring that health care markets operate well, but they are not alone. The functioning of health care markets is also profoundly affected by other parts of the federal government (notably the Centers for Medicare and Medicaid Services) and by state legislation and regulation. In this current period of such dynamic change, it is particularly important for the antitrust agencies to continue and enhance their communication and coordination with other government agencies as well as to maintain vigilant antitrust enforcement and consumer protection in health care markets. © 2014 Project HOPE-The People-to-People Health Foundation, Inc.


Anna S.L.,Carnegie Mellon University
Annual Review of Fluid Mechanics | Year: 2016

Precise, tunable emulsions and foams produced in microfluidic geometries have found wide application in biochemical analysis and materials synthesis and characterization. Superb control of the volume, uniformity, and generation rate of droplets and bubbles arises from unique features of the microscale behavior of fluid interfaces. Fluid interfaces confined within microfluidic channels behave quite differently than their counterparts in unbounded flows. Confinement inhibits capillary instabilities so that breakup occurs by largely quasi-static mechanisms. The three-dimensional flow near confined interfaces in rectangular geometries and feedback effects from resistance changes in the entire microfluidic network play important roles in regulating the interfacial deformation. Timescales for transport of surfactants and particles to interfaces compete with flow timescales at the microscale, providing further opportunity for tuning the interfacial coverage and properties of individual droplets and bubbles. © Copyright 2016 by Annual Reviews. All rights reserved.


De La Cruz J.,University of Seville | Karbstein K.,Scripps Research Institute | Woolford J.L.,Carnegie Mellon University
Annual Review of Biochemistry | Year: 2015

The proteome of cells is synthesized by ribosomes, complex ribonucleoproteins that in eukaryotes contain 79-80 proteins and four ribosomal RNAs (rRNAs) more than 5,400 nucleotides long. How these molecules assemble together and how their assembly is regulated in concert with the growth and proliferation of cells remain important unanswered questions. Here, we review recently emerging principles to understand how eukaryotic ribosomal proteins drive ribosome assembly in vivo. Most ribosomal proteins assemble with rRNA cotranscriptionally; their association with nascent particles is strengthened as assembly proceeds. Each subunit is assembled hierarchically by sequential stabilization of their subdomains. The active sites of both subunits are constructed last, perhaps to prevent premature engagement of immature ribosomes with active subunits. Late-assembly intermediates undergo quality-control checks for proper function. Mutations in ribosomal proteins that affect mostly late steps lead to ribosomopathies, diseases that include a spectrum of cell type-specific disorders that often transition from hypoproliferative to hyperproliferative growth. Copyright © 2015 by Annual Reviews. All rights reserved.


Manadhata P.K.,Symantec | Wing J.M.,Carnegie Mellon University
IEEE Transactions on Software Engineering | Year: 2011

Measurement of software security is a long-standing challenge to the research community. At the same time, practical security metrics and measurements are essential for secure software development. Hence, the need for metrics is more pressing now due to a growing demand for secure software. In this paper, we propose using a software system's attack surface measurement as an indicator of the system's security. We formalize the notion of a system's attack surface and introduce an attack surface metric to measure the attack surface in a systematic manner. Our measurement method is agnostic to a software system's implementation language and is applicable to systems of all sizes; we demonstrate our method by measuring the attack surfaces of small desktop applications and large enterprise systems implemented in C and Java. We conducted three exploratory empirical studies to validate our method. Software developers can mitigate their software's security risk by measuring and reducing their software's attack surfaces. Our attack surface reduction approach complements the software industry's traditional code quality improvement approach for security risk mitigation and is useful in multiple phases of the software development lifecycle. Our collaboration with SAP demonstrates the use of our metric in the software development process. © 2011 IEEE Published by the IEEE Computer Society.


Spirtes P.,Carnegie Mellon University
Journal of Machine Learning Research | Year: 2010

The goal of many sciences is to understand the mechanisms by which variables came to take on the values they have (that is, to find a generative model), and to predict what the values of those variables would be if the naturally occurring mechanisms were subject to outside manipulations. The past 30 years has seen a number of conceptual developments that are partial solutions to the problem of causal inference from observational sample data or a mixture of observational sample and experimental data, particularly in the area of graphical causal modeling. However, in many domains, problems such as the large numbers of variables, small samples sizes, and possible presence of unmeasured causes, remain serious impediments to practical applications of these developments. The articles in the Special Topic on Causality address these and other problems in applying graphical causal modeling algorithms. This introduction to the Special Topic on Causality provides a brief introduction to graphical causal modeling, places the articles in a broader context, and describes the differences between causal inference and ordinary machine learning classification and prediction problems. © 2010 Peter Spirtes.


Cohn J.,Carnegie Mellon University
IEEE Signal Processing Magazine | Year: 2010

The face conveys information about a person's age, sex, background, and identity; what they are feeling, thinking, or likely to do next. Facial expression regulates face-to-face interactions, indicates reciprocity and interpersonal attraction or repulsion, and enables intersubjectivity between members of different cultures. Facial expression indexes neurological and psychiatric functioning and reveals personality and socioemotional development. Not surprisingly, the face has been of keen interest to behavioral scientists. © 2010 IEEE.


Roth A.,Carnegie Mellon University | Roughgarden T.,Stanford University
Proceedings of the Annual ACM Symposium on Theory of Computing | Year: 2010

We define a new interactive differentially private mechanism - the median mechanism - for answering arbitrary predicate queries that arrive online. Given fixed accuracy and privacy constraints, this mechanism can answer exponentially more queries than the previously best known interactive privacy mechanism (the Laplace mechanism, which independently perturbs each query result). With respect to the number of queries, our guarantee is close to the best possible, even for non-interactive privacy mechanisms. Conceptually, the median mechanism is the first privacy mechanism capable of identifying and exploiting correlations among queries in an interactive setting. We also give an efficient implementation of the median mechanism, with running time polynomial in the number of queries, the database size, and the domain size. This efficient implementation guarantees privacy for all input databases, and accurate query results for almost all input distributions. The dependence of the privacy on the number of queries in this mechanism improves over that of the best previously known efficient mechanism by a super-polynomial factor, even in the non-interactive setting. © 2010 ACM.


Anwar S.,Carnegie Mellon University | Loughran T.A.,University of Maryland University College
Criminology | Year: 2011

The effect of criminal experience on risk perceptions is of central importance to deterrence theory but has been vastly understudied. This article develops a realistic Bayesian learning model of how individuals will update their risk perceptions over time in response to the signals they receive during their offending experiences. This model implies a simple function that we estimate to determine the deterrent effect of an arrest. We find that an individual who commits one crime and is arrested will increase his or her perceived probability of being caught by 6.3 percent compared with if he or she had not been arrested. We also find evidence that the more informative the signal received by an individual is, the more he or she will respond to it, which is consistent with more experienced offenders responding less to an arrest than less experienced offenders do. Parsing our results out by type of crime indicates that an individual who is arrested for an aggressive crime will increase both his or her aggressive crime risk perception as well as his or her income-generating crime risk perception, although the magnitude of the former may be slightly larger. This implies that risk perception updating, and thus potentially deterrence, may be partially, although not completely, crime specific. © 2011 American Society of Criminology.


Collins S.H.,Carnegie Mellon University | Bruce Wiggin M.,University of North Carolina at Chapel Hill | Sawicki G.S.,University of North Carolina at Chapel Hill
Nature | Year: 2015

With efficiencies derived from evolution, growth and learning, humans are very well-tuned for locomotion. Metabolic energy used during walking can be partly replaced by power input from an exoskeleton, but is it possible to reduce metabolic rate without providing an additional energy source? This would require an improvement in the efficiency of the human-machine system as a whole, and would be remarkable given the apparent optimality of human gait. Here we show that the metabolic rate of human walking can be reduced by an unpowered ankle exoskeleton. We built a lightweight elastic device that acts in parallel with the user's calf muscles, off-loading muscle force and thereby reducing the metabolic energy consumed in contractions. The device uses a mechanical clutch to hold a spring as it is stretched and relaxed by ankle movements when the foot is on the ground, helping to fulfil one function of the calf muscles and Achilles tendon. Unlike muscles, however, the clutch sustains force passively. The exoskeleton consumes no chemical or electrical energy and delivers no net positive mechanical work, yet reduces the metabolic cost of walking by 7.2 ± 2.6% for healthy human users under natural conditions, comparable to savings with powered devices. Improving upon walking economy in this way is analogous to altering the structure of the body such that it is more energy-effective at walking. While strong natural pressures have already shaped human locomotion, improvements in efficiency are still possible. Much remains to be learned about this seemingly simple behaviour. © 2015 Macmillan Publishers Limited. All rights reserved.


He J.,Carnegie Mellon University
Nature Methods | Year: 2016

Upon illumination, photosensitizer molecules produce reactive oxygen species that can be used for functional manipulation of living cells, including protein inactivation, targeted-damage introduction and cellular ablation. Photosensitizers used to date have been either exogenous, resulting in delivery and removal challenges, or genetically encoded proteins that form or bind a native photosensitizing molecule, resulting in a constitutively active photosensitizer inside the cell. We describe a genetically encoded fluorogen-activating protein (FAP) that binds a heavy atom−substituted fluorogenic dye, forming an 'on-demand' activated photosensitizer that produces singlet oxygen and fluorescence when activated with near-infrared light. This targeted and activated photosensitizer (TAPs) approach enables protein inactivation, targeted cell killing and rapid targeted lineage ablation in living larval and adult zebrafish. The near-infrared excitation and emission of this FAP-TAPs provides a new spectral range for photosensitizer proteins that could be useful for imaging, manipulation and cellular ablation deep within living organisms. © 2016 Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved.


Melo F.S.,GAIPS | Veloso M.,Carnegie Mellon University
Artificial Intelligence | Year: 2011

Creating coordinated multiagent policies in environments with uncertainty is a challenging problem, which can be greatly simplified if the coordination needs are known to be limited to specific parts of the state space. In this work, we explore how such local interactions can simplify coordination in multiagent systems. We focus on problems in which the interaction between the agents is sparse and contribute a new decision-theoretic model for decentralized sparse-interaction multiagent systems, Dec-SIMDPs, that explicitly distinguishes the situations in which the agents in the team must coordinate from those in which they can act independently. We relate our new model to other existing models such as MMDPs and Dec-MDPs. We then propose a solution method that takes advantage of the particular structure of Dec-SIMDPs and provide theoretical error bounds on the quality of the obtained solution. Finally, we show a reinforcement learning algorithm in which independent agents learn both individual policies and when and how to coordinate. We illustrate the application of the algorithms throughout the paper in several multiagent navigation scenarios. © 2011 Elsevier B.V.


Kalidindi S.R.,Georgia Institute of Technology | De Graef M.,Carnegie Mellon University
Annual Review of Materials Research | Year: 2015

The field of materials science and engineering is on the cusp of a digital data revolution. After reviewing the nature of data science and Big Data, we discuss the features of materials data that distinguish them from data in other fields. We introduce the concept of process-structure-property (PSP) linkages and illustrate how the determination of PSPs is one of the main objectives of materials data science. Then we review a selection of materials databases, as well as important aspects of materials data management, such as storage hardware, archiving strategies, and data access strategies. We introduce the emerging field of materials data analytics, which focuses on data-driven approaches to extract and curate materials knowledge from available data sets. The critical need for materials e-collaboration platforms is highlighted, and we conclude the article with a number of suggestions regarding the near-term future of the materials data science field. Copyright © 2015 by Annual Reviews. All rights reserved.


Brown K.W.,Virginia Commonwealth University | Weinstein N.,University of Essex | Creswell J.D.,Carnegie Mellon University
Psychoneuroendocrinology | Year: 2012

Background: Individual differences in mindfulness have been associated with numerous self-report indicators of stress, but research has not examined how mindfulness may buffer neuroendocrine and psychological stress responses under controlled laboratory conditions. The present study investigated the role of trait mindfulness in buffering cortisol and affective responses to a social evaluative stress challenge versus a control task. Methods: Participants completed measures of trait mindfulness, perceived stress, anxiety, and fear of negative evaluation before being randomized to complete the Trier Social Stress Test (TSST; Kirschbaum et al., 1993) or a control task. At points throughout the session, participants provided five saliva samples to assess cortisol response patterns, and completed four self-report measures of anxiety and negative affect to assess psychological responses. Results: In accord with hypotheses, higher trait mindfulness predicted lower cortisol responses to the TSST, relative to the control task, as well as lower anxiety and negative affect. These relations remained significant when controlling for the role of other variables that predicted cortisol and affective responses. Conclusions: The findings suggest that trait mindfulness modulates cortisol and affective responses to an acute social stressor. Further research is needed to understand the neural pathways through which mindfulness impacts these responses. © 2012 Elsevier Ltd.


Khair A.S.,Carnegie Mellon University
Journal of Colloid and Interface Science | Year: 2012

We quantify the phoretic migration of a spherical cation-permselective colloidal particle immersed in a binary electrolyte under a time-dependent electric field. We invoke the thin-Debye-layer approximation, where the size of ionic Debye layer enveloping the particle is much smaller than the particle radius. The imposed electric field generates ion concentration gradients, or concentration polarization, in the bulk (electroneutral) electrolyte outside the Debye layer. The bulk ion concentration polarization-and consequently the particle's phoretic velocity-evolves on the time scale for ion diffusion around the particle, which can be on the order of milliseconds for typical colloidal dimensions. Notably, concentration polarization arises here solely due to the permselectivity of the particle; it does not require non-uniform ionic transport in the Debye layer (i.e., surface conduction). Thus, the phoretic transport of a permselective particle is significantly different to that of a inert, dielectric particle, since surface conduction is necessary to achieve bulk concentration polarization in the (more commonly studied) latter case. Calculations are presented for a permselective particle under oscillatory (ac) and suddenly applied electric fields. In the former case, the particle velocity possesses frequency-dependent components in phase and out of phase with the driving field; in the latter case, the particle approaches its terminal velocity with a long-time (algebraic) tail. © 2012 Elsevier Inc.


Zollman K.J.,Carnegie Mellon University
Proceedings. Biological sciences / The Royal Society | Year: 2013

Costly signalling theory has become a common explanation for honest communication when interests conflict. In this paper, we provide an alternative explanation for partially honest communication that does not require significant signal costs. We show that this alternative is at least as plausible as traditional costly signalling, and we suggest a number of experiments that might be used to distinguish the two theories.


Nagin D.S.,Carnegie Mellon University | Snodgrass G.M.,State of Alaska
Journal of Quantitative Criminology | Year: 2013

Objectives: This paper uses a sample of convicted offenders from Pennsylvania to estimate the effect of incarceration on post-release criminality. Methods: To do so, we capitalize on a feature of the criminal justice system in Pennsylvania-the county-level randomization of cases to judges. We begin by identifying five counties in which there is substantial variation across judges in the uses of incarceration, but no evidence indicating that the randomization process had failed. The estimated effect of incarceration on rearrest is based on comparison of the rearrest rates of the caseloads of judges with different proclivities for the use of incarceration. Results: Using judge as an instrumental variable, we estimate a series of confidence intervals for the effect of incarceration on one year, two year, five year, and ten year rearrest rates. Conclusions: On the whole, there is little evidence in our data that incarceration impacts rearrest. © 2013 Springer Science+Business Media New York.


Rohrer G.S.,Carnegie Mellon University
Journal of Materials Science | Year: 2011

This paper reviews findings on the anisotropy of the grain boundary energies. After introducing the basic concepts, there is a discussion of fundamental models used to understand and predict grain boundary energy anisotropy. Experimental methods for measuring the grain boundary energy anisotropy, all of which involve application of the Herring equation, are then briefly described. The next section reviews and compares the results of measurements and model calculations with the goal of identifying generally applicable characteristics. This is followed by a brief discussion of the role of grain boundary energies in nucleating discontinuous transitions in grain boundary structure and chemistry, known as complexion transitions. The review ends with some questions to be addressed by future research and a summary of what is known about grain boundary energy anisotropy. © Springer Science+Business Media, LLC 2011.


Platzer A.,Carnegie Mellon University
HSCC'11 - Proceedings of the 2011 ACM/SIGBED Hybrid Systems: Computation and Control | Year: 2011

We address the verification problem for distributed hybrid systems with nontrivial dynamics. Consider air traffic collision avoidance maneuvers, for example. Verifying dynamic appearance of aircraft during an ongoing collision avoidance maneuver is a longstanding and essentially unsolved problem. The resulting systems are not hybrid systems and their state space is not of the form ℝn. They are distributed hybrid systems with nontrivial continuous and discrete dynamics in distributed state spaces whose dimension and topology changes dynamically over time. We present the first formal verification technique that can handle the complicated nonlinear dynamics of these systems. We introduce quantified differential invariants, which are properties that can be checked for invariance along the dynamics of the distributed hybrid system based on differentiation, quantified substitution, and quantifier elimination in real-closed fields. This gives a computationally attractive technique, because it works without having to solve the infinite-dimensional differential equation systems underlying distributed hybrid systems. We formally verify a roundabout maneuver in which aircraft can appear dynamically. Copyright 2011 ACM.


Fienberg S.E.,Carnegie Mellon University
Statistical Science | Year: 2011

Starting with the neo-Bayesian revival of the 1950s, many statisticians argued that it was inappropriate to use Bayesian methods, and in particular subjective Bayesian methods in governmental and public policy settings because of their reliance upon prior distributions. But the Bayesian framework often provides the primary way to respond to questions raised in these settings and the numbers and diversity of Bayesian applications have grown dramatically in recent years. Through a series of examples, both historical and recent, we argue that Bayesian approaches with formal and informal assessments of priors AND likelihood functions are well accepted and should become the norm in public settings. Our examples include censustaking and small area estimation, US election night forecasting, studies reported to the US Food and Drug Administration, assessing global climate change, and measuring potential declines in disability among the elderly. © Institute of Mathematical Statistics, 2011.


Griffiths R.B.,Carnegie Mellon University
Foundations of Physics | Year: 2014

It is shown how all the major conceptual difficulties of standard (textbook) quantum mechanics, including the two measurement problems and the (supposed) nonlocality that conflicts with special relativity, are resolved in the consistent or decoherent histories interpretation of quantum mechanics by using a modified form of quantum logic to discuss quantum properties (subspaces of the quantum Hilbert space), and treating quantum time development as a stochastic process. The histories approach in turn gives rise to some conceptual difficulties, in particular the correct choice of a framework (probabilistic sample space) or family of histories, and these are discussed. The central issue is that the principle of unicity, the idea that there is a unique single true description of the world, is incompatible with our current understanding of quantum mechanics. © 2014 Springer Science+Business Media New York.


Rubin E.S.,Carnegie Mellon University
International Journal of Greenhouse Gas Control | Year: 2012

This paper reviews and compares the prevailing methods, metrics and assumptions underlying cost estimates for CO 2 capture and storage (CCS) technologies applied to fossil fuel power plants. This assessment reveals a number of significant differences and inconsistencies across different studies, not only in key technical, economic and financial assumptions related to the cost of a CCS project (such as differences in plant size, fuel type, capacity factor, and cost of capital) but also in the underlying methods and cost elements that are included (or excluded) in a particular study (such as the omission of certain "owner's" costs or the cost of transport and storage). Such differences often are not apparent in the cost results that are reported publicly or in the technical literature. In other cases, measures that have very different meanings (such as the costs of CO 2 avoided, CO 2 captured and CO 2 abated) are all reported in similar units of "dollars per ton CO 2" As a consequence, there is likely to be some degree confusion, misunderstanding and possible mis-representation of CCS costs. Given the widespread interest in the cost of CCS and the potential for lower-cost CO 2 capture technology, methods to improve the consistency and transparency of CCS cost estimates are needed. © 2012 Elsevier Ltd.


Mukhopadhyay S.,University of Texas at Austin | Linstedt A.D.,Carnegie Mellon University
Journal of Molecular Medicine | Year: 2013

Bacterial AB5 toxins are a clinically relevant class of exotoxins that include several well-known members such as Shiga, cholera, and pertussis toxins. Infections with toxin-producing bacteria cause devastating human diseases that affect millions of individuals each year and have no definitive medical treatment. The molecular targets of AB5 toxins reside in the cytosol of infected cells, and the toxins reach the cytosol by trafficking through the retrograde membrane transport pathway that avoids degradative late endosomes and lysosomes. Focusing on Shiga toxin as the archetype member, we review recent advances in understanding the molecular mechanisms involved in the retrograde trafficking of AB5 toxins and highlight how these basic science advances are leading to the development of a promising new therapeutic approach based on inhibiting toxin transport. © 2013 Springer-Verlag Berlin Heidelberg.


Sekerka R.F.,Carnegie Mellon University
Philosophical Magazine | Year: 2011

We develop the irreversible thermodynamic basis of the phase field model, which is a mesoscopic diffuse interface model that eliminates interface tracking during phase transformations. The phase field is an auxiliary parameter that identifies the phase; it is continuous but makes a transition over a thin region, the diffuse interface, from its constant value in a growing phase to some other value in the nutrient phase. All phases are treated thermodynamically as viscous liquids, even crystalline solids. Phases are assumed to be isotropic for simplicity with reference to works that include anisotropy. The basis is an entropy functional which is an integral of an entropy density that includes non-classical gradient entropies. Equilibrium is investigated to identify a non-classical temperature and non-classical chemical potentials for a multicomponent system that are uniform at equilibrium in the absence of external forces. Coupled partial differential equations that govern the time evolution of the phase field and accompanying fields (such as temperature and composition) are formulated on the basis of local positive entropy production subject to suitable constraints on energy and chemical species. Fluxes of energy and chemical species, Korteweg stresses due to inhomogeneities and an equation for phase field evolution are obtained from the rate of entropy production by postulating linear constitutive relations. The full phase field equations are discussed and illustrated for the simple case of solidification of a pure material from its melt assuming uniform density. © 2011 Taylor & Francis.


Platzer A.,Carnegie Mellon University
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2010

We address a fundamental mismatch between the combinations of dynamics that occur in complex physical systems and the limited kinds of dynamics supported in analysis. Modern applications combine communication, computation, and control. They may even form dynamic networks, where neither structure nor dimension stay the same while the system follows mixed discrete and continuous dynamics. We provide the logical foundations for closing this analytic gap. We develop a system model for distributed hybrid systems that combines quantified differential equations with quantified assignments and dynamic dimensionality-changes. We introduce a dynamic logic for verifying distributed hybrid systems and present a proof calculus for it. We prove that this calculus is a sound and complete axiomatization of the behavior of distributed hybrid systems relative to quantified differential equations. In our calculus we have proven collision freedom in distributed car control even when new cars may appear dynamically on the road. © 2010 Springer-Verlag Berlin Heidelberg.


Griffiths R.B.,Carnegie Mellon University
Foundations of Physics | Year: 2011

It is argued that while quantum mechanics contains nonlocal or entangled states, the instantaneous or nonlocal influences sometimes thought to be present due to violations of Bell inequalities in fact arise from mistaken attempts to apply classical concepts and introduce probabilities in a manner inconsistent with the Hilbert space structure of standard quantum mechanics. Instead, Einstein locality is a valid quantum principle: objective properties of individual quantum systems do not change when something is done to another noninteracting system. There is no reason to suspect any conflict between quantum theory and special relativity. © 2010 Springer Science+Business Media, LLC.


Cheraghchi M.,Carnegie Mellon University
Algorithmica | Year: 2013

The basic goal in combinatorial group testing is to identify a set of up to d defective items within a large population of size n≫d using a pooling strategy. Namely, the items can be grouped together in pools, and a single measurement would reveal whether there are one or more defectives in the pool. The threshold model is a generalization of this idea where a measurement returns positive if the number of defectives in the pool reaches a fixed threshold u>0, negative if this number is no more than a fixed lower threshold ℓ0. © 2013 Springer Science+Business Media New York.


Yaan O.,Carnegie Mellon University
Physical Review E - Statistical, Nonlinear, and Soft Matter Physics | Year: 2015

We consider a power system with N transmission lines whose initial loads (i.e., power flows) L1,⋯,LN are independent and identically distributed with PL(x)=PL≤x. The capacity Ci defines the maximum flow allowed on line i and is assumed to be given by Ci=(1+α)Li, with α>0. We study the robustness of this power system against random attacks (or failures) that target a p fraction of the lines, under a democratic fiber-bundle-like model. Namely, when a line fails, the load it was carrying is redistributed equally among the remaining lines. Our contributions are as follows. (i) We show analytically that the final breakdown of the system always takes place through a first-order transition at the critical attack size p∗=1-ELmaxx(PL>x(αx+EL|L>x)), where E· is the expectation operator; (ii) we derive conditions on the distribution PL(x) for which the first-order breakdown of the system occurs abruptly without any preceding diverging rate of failure; (iii) we provide a detailed analysis of the robustness of the system under three specific load distributions - uniform, Pareto, and Weibull - showing that with the minimum load Lmin and mean load EL fixed, Pareto distribution is the worst (in terms of robustness) among the three, whereas Weibull distribution is the best with shape parameter selected relatively large; (iv) we provide numerical results that confirm our mean-field analysis; and (v) we show that p∗ is maximized when the load distribution is a Dirac delta function centered at EL, i.e., when all lines carry the same load. This last finding is particularly surprising given that heterogeneity is known to lead to high robustness against random failures in many other systems. © 2015 American Physical Society.


Mandelbaum R.,Carnegie Mellon University
Journal of Instrumentation | Year: 2015

We present a pedagogical review of the weak gravitational lensing measurement process and its connection to major scientific questions such as dark matter and dark energy. Then we describe common ways of parametrizing systematic errors and understanding how they affect weak lensing measurements. Finally, we discuss several instrumental systematics and how they fit into this context, and conclude with some future perspective on how progress can be made in understanding the impact of instrumental systematics on weak lensing measurements. © 2015 IOP Publishing Ltd and Sissa Medialab srl.


Swendsen R.H.,Carnegie Mellon University
Physical Review E - Statistical, Nonlinear, and Soft Matter Physics | Year: 2015

The proper definition of entropy is fundamental to the relationship between statistical mechanics and thermodynamics. It also plays a major role in the recent debate about the validity of the concept of negative temperature. In this paper, I analyze and calculate the thermodynamic entropy for large but finite quantum mechanical systems. A special feature of this analysis is that the thermodynamic energy of a quantum system is shown to be a continuous variable, rather than being associated with discrete energy eigenvalues. Calculations of the entropy as a function of energy can be carried out with a Legendre transform of thermodynamic potentials obtained from a canonical ensemble. The resultant expressions for the entropy are able to describe equilibrium between quantum systems having incommensurate energy-level spacings. This definition of entropy preserves all required thermodynamic properties, including satisfaction of all postulates and laws of thermodynamics. It demonstrates the consistency of the concept of negative temperature with the principles of thermodynamics. © 2015 American Physical Society.


Bruchez M.P.,Carnegie Mellon University
Current Opinion in Chemical Biology | Year: 2011

Thirteen years after the demonstration of quantum dots as biological imaging agents, and nine years after the initial commercial introduction of bioconjugated quantum dots, the brightness and photostability of the quantum dots has enabled a range of investigations using single molecule tracking. These materials are being routinely utilized by a number of groups to track the dynamics of single molecules in reconstituted biophysical systems and on living cells, and are especially powerful for investigations of single molecules over long timescales with short exposure times and high pointing accuracy. New approaches are emerging where the quantum dots are used as 'hard-sphere' probes for intracellular compartments. Innovations in quantum dot surface modification are poised to substantially expand the utility of these materials. © 2011 Elsevier Ltd.


Jirout J.,Temple University | Klahr D.,Carnegie Mellon University
Developmental Review | Year: 2012

Although curiosity is an undeniably important aspect of children's cognitive development, a universally accepted operational definition of children's curiosity does not exist. Almost all of the research on measuring curiosity has focused on adults, and has used predominately questionnaire-type measures that are not appropriate for young children. In this review we (a) synthesize the range of definitions and measures of children's curiosity and (b) propose a new operational definition and measurement procedure for assessing and advancing scientific curiosity in young children. In the first part of the paper, we summarize Loewenstein's (1994) review of theoretical perspectives on adult curiosity, and critically evaluate a wide range of efforts to create and implement operational measures of curiosity, focusing mainly on behavioral measures of curiosity in children. In the second part, we return to Loewenstein's theory and present an argument for adopting his " information-gap" theory of curiosity as a framework for reviewing various procedures that have been suggested for measuring children's exploratory curiosity. Finally, we describe a new paradigm for measuring exploratory curiosity in preschool children, defining curiosity as the threshold of desired uncertainty in the environment that leads to exploratory behavior. We present data demonstrating the reliability and validity of this measure, discuss initial results on developmental differences in young children's curiosity, and conclude with a general summary and suggestions for future research. © 2012 Elsevier Inc..


Armitage B.A.,Carnegie Mellon University
Current Opinion in Chemical Biology | Year: 2011

Fluorescence microscopy and molecular tagging technologies have ushered in a new era in our understanding of protein localization and function in cells. This review summarizes recent efforts to extend some of these methods (and to create new ones) to imaging of RNA in live cells. Both fluorescent proteins and hybridization probes allow noncovalent labeling of specific RNA molecules with fluorescent dyes that allow detection and tracking in real time. © 2011 Elsevier Ltd.


This paper explores the role of the state in re-architecting social networks and thereby new technology directions in the United States. It draws on a case study of DARPA's Microsystems Technology Office from 1992 to 2008. Leveraging one of the most radical directorships in DARPA's history, I argue that the perceived "death" of DARPA under Tony Tether was because past analyses, by focusing on the organization's culture and structure, overlooked a set of lasting, informal institutions among DARPA program managers. I find that despite significant changes in the recipients and outcomes of DARPA attentions, these same institutions for directing technology were in place both before and during Tether's directorship. Drawing on these results, I suggest that we must add to technology policy-making a new option - embedded network governance. © 2010 Elsevier B.V. All rights reserved.


Yeh S.,University of California at Davis | Rubin E.S.,Carnegie Mellon University
Energy Economics | Year: 2012

The use of log-linear experience curves (or learning curves) relating reductions in the unit cost of technologies to their cumulative production or installed capacity has become a common method of representing endogenous technical change in energy-economic models used for policy analysis. Yet, there are significant uncertainties in such formulations whose impact on key model results have been insufficiently examined or considered. This paper reviews the major types of uncertainty in log-linear experience curves and their effect on projected rates of cost reduction. Uncertainties are found not only in the learning rate parameter of a log-linear model, but also in the functional form that determines the shape of an experience curve. Evidence for alternative forms such as an S-shaped curve is reviewed along with case studies that demonstrate the uncertainties associated with cost increases during early commercialization of a technology-a phenomena that is widely recognized but rarely quantified or incorporated in learning models. Additional factors discussed include the effects of learning discontinuities, institutional forgetting, and the influence of social, economic and political factors. We then review other models of causality, which aim to improve modelers' ability to explain and predict the influence of other underlying processes that contribute to technology cost reductions in addition to learning. Ignoring other types of underlying mechanisms can create a false sense of precision and overestimate the true contribution of learning. Currently, however, uncertainties in such multi-factor models remain large due to the difficulties of estimating key parameters (such as private-sector R&D investments) and extending models of a specific technology to a broader suite of technologies and cost projections. Pending the development and validation of more robust models of technological change, we suggest ways to significantly improve the characterization and reporting of current learning model uncertainties and their impacts on the results of energy-economic models to help reduce the potential for drawing inappropriate or erroneous policy conclusions. © 2011 Elsevier B.V.


Mengshoel O.J.,Carnegie Mellon University
Artificial Intelligence | Year: 2010

One of the main approaches to performing computation in Bayesian networks (BNs) is clique tree clustering and propagation. The clique tree approach consists of propagation in a clique tree compiled from a BN, and while it was introduced in the 1980s, there is still a lack of understanding of how clique tree computation time depends on variations in BN size and structure. In this article, we improve this understanding by developing an approach to characterizing clique tree growth as a function of parameters that can be computed in polynomial time from BNs, specifically: (i) the ratio of the number of a BN's non-root nodes to the number of root nodes, and (ii) the expected number of moral edges in their moral graphs. Analytically, we partition the set of cliques in a clique tree into different sets, and introduce a growth curve for the total size of each set. For the special case of bipartite BNs, there are two sets and two growth curves, a mixed clique growth curve and a root clique growth curve. In experiments, where random bipartite BNs generated using the BPART algorithm are studied, we systematically increase the out-degree of the root nodes in bipartite Bayesian networks, by increasing the number of leaf nodes. Surprisingly, root clique growth is well-approximated by Gompertz growth curves, an S-shaped family of curves that has previously been used to describe growth processes in biology, medicine, and neuroscience. We believe that this research improves the understanding of the scaling behavior of clique tree clustering for a certain class of Bayesian networks; presents an aid for trade-off studies of clique tree clustering using growth curves; and ultimately provides a foundation for benchmarking and developing improved BN inference and machine learning algorithms. © 2010 Elsevier B.V. All rights reserved.


Singh P.V.,Carnegie Mellon University
ACM Transactions on Software Engineering and Methodology | Year: 2010

In this study we investigate the impact of community-level networksrelationships that exist among developers in an OSS communityon the productivity of member developers. Specifically, we argue that OSS community networks characterized by small-world properties would positively influence the productivity of the member developers by providing them with speedy and reliable access to more quantity and variety of information and knowledge resources. Specific hypotheses are developed and tested using longitudinal data on a large panel of 4,279 projects from 15 different OSS communities hosted at Sourceforge. Our results suggest that significant variation exists in small-world properties of OSS communities at Sourceforge. After accounting for project, foundry, and time-specific observed and unobserved effects, we found a statistically significant relationship between small-world properties of a community and the technical and commercial success of the software produced by its members. In contrast to the findings of prior research, we also found the lack of a significant relationship between closeness and betweenness centralities of the project teams and their success. These results were robust to a number of controls and model specifications. © 2010 ACM.


Thiessen E.D.,Carnegie Mellon University
Cognitive Science | Year: 2010

Infant and adult learners are able to identify word boundaries in fluent speech using statistical information. Similarly, learners are able to use statistical information to identify word-object associations. Successful language learning requires both feats. In this series of experiments, we presented adults and infants with audio-visual input from which it was possible to identify both word boundaries and word-object relations. Adult learners were able to identify both kinds of statistical relations from the same input. Moreover, their learning was actually facilitated by the presence of two simultaneously present relations. Eight-month-old infants, however, do not appear to benefit from the presence of regular relations between words and objects. Adults, like 8-month-olds, did not benefit from regular audio-visual correspondences when they were tested with tones, rather than linguistic input. These differences in learning outcomes across age and input suggest that both developmental and stimulus-based constraints affect statistical learning. © 2010 Cognitive Science Society, Inc.


Helgeson V.S.,Carnegie Mellon University
Psycho-Oncology | Year: 2011

Objective: The goal of this research was to examine the extent to which 10-year breast cancer survivors integrated cancer into their self-concept (i.e. survivor centrality), identify predictors of survivor centrality, and determine the relation of survivor centrality to well-being. Methods: Breast cancer survivors (n = 240) were interviewed 10 years following the initial diagnosis. They completed measures of survivor centrality, illness valence (i.e. positive or negative views of illness), and well-being (positive and negative affect, mental and physical functioning, psychological distress, benefit finding). Results: There were few predictors of the kinds of women who were more likely to integrate breast cancer into their self-concepts, but survivor centrality was related to engaging in behaviors that suggested survivorship was relevant to women's daily lives, such as becoming involved in breast cancer activities. Survivor centrality was related to three markers of negative psychological well-being: more negative affect, poorer mental functioning, and greater psychological distress. However, in the case of negative affect and psychological distress, this relation was moderated by illness valence, such that survivor centrality was only related to negative psychological well-being when the illness was viewed in less positive terms. Conclusions: Women vary in the extent to which they define themselves in terms of the breast cancer experience. Survivor centrality in and of itself is not always indicative of adjustment to disease. When women have a more negative view of being a breast cancer survivor, survivor centrality is more likely to signify potential problems. Copyright © 2010 John Wiley & Sons, Ltd.


Nagin D.S.,Carnegie Mellon University | Odgers C.L.,University of California at Irvine
Annual Review of Clinical Psychology | Year: 2010

Group-based trajectory models are increasingly being applied in clinical research to map the developmental course of symptoms and assess heterogeneity in response to clinical interventions. In this review, we provide a nontechnical overview of group-based trajectory and growth mixture modeling alongside a sampling of how these models have been applied in clinical research. We discuss the challenges associated with the application of both types of group-based models and propose a set of preliminary guidelines for applied researchers to follow when reporting model results. Future directions in group-based modeling applications are discussed, including the use of trajectory models to facilitate causal inference when random assignment to treatment condition is not possible. Copyright © 2010 by Annual Reviews. All rights reserved.


Kim M.,Carnegie Mellon University
Pattern Recognition | Year: 2010

We tackle the structured output classification problem using the Conditional Random Fields (CRFs). Unlike the standard 0/1 loss case, we consider a cost-sensitive learning setting where we are given a non-0/1 misclassification cost matrix at the individual output level. Although the task of cost-sensitive classification has many interesting practical applications that retain domain-specific scales in the output space (e.g., hierarchical or ordinal scale), most CRF learning algorithms are unable to effectively deal with the cost-sensitive scenarios as they merely assume a nominal scale (hence 0/1 loss) in the output space. In this paper, we incorporate the cost-sensitive loss into the large margin learning framework. By large margin learning, the proposed algorithm inherits most benefits from the SVM-like margin-based classifiers, such as the provable generalization error bounds. Moreover, the soft-max approximation employed in our approach yields a convex optimization similar to the standard CRF learning with only slight modification in the potential functions. We also provide the theoretical cost-sensitive generalization error bound. We demonstrate the improved prediction performance of the proposed method over the existing approaches in a diverse set of sequence/image structured prediction problems that often arise in pattern recognition and computer vision domains. © 2010 Elsevier Ltd. All rights reserved.


The passage of ionic current across a charge-selective surface has been studied for over a century and is relevant to well-established processes such as electrodialysis, electrodeposition, and electrochromatography. Recent years have witnessed a resurgence of interest in this subject, motivated by experiments demonstrating charge-selective transport of ions and solutes in nanofluidic devices. In this paper, we revisit and build upon the prototypical problem of one-dimensional ion transport across a flat ideally ion-selective surface, by examining the influence of imposed fluid flows on concentration polarization, over-limiting current, and second-kind (non-equilibrium) electro-osmotic instability at the surface. Specifically, we consider a simple model system of a cation-selective surface or membrane that admits a uniform fluid flow across itself. The membrane resides against a binary symmetric electrolyte, whose concentration is uniform in a "well-mixed" region at a prescribed distance from the membrane. A potential difference across the system drives an ionic current, leading to concentration polarization in the "unstirred layer" between the membrane and well-mixed bulk. The concentration polarization profile reflects a balance between advection of ions with the imposed "normal flow" and diffusion. The relative importance of these effects is parameterized by a Pećlet number Pe; notably, Pe is a signed quantity as the flow can be imposed toward or away from the membrane. An asymptotic analysis in the thin-Debye-layer limit reveals a significant impact of normal flow on concentration polarization and the advection-diffusion limiting current across the membrane. In particular, there exists a nonlinear concentration profile in the unstirred layer for non-zero Pe, in contrast to the familiar linear (diffusive) concentration polarization at Pe = 0. Next, we use matched asymptotic expansions to explore the structure of the unstirred layer at over-limiting currents, wherein a non-equilibrium space-charge layer develops near the membrane surface. A key step in this process is the derivation of a "generalized master equation" for the electric field across the unstirred layer. Finally, we examine the instability of the quiescent concentration polarization resulting from second-kind electro-osmotic slip in the space-charge layer. A linear stability analysis shows that normal flow can either enhance or retard the instability, depending on the flow direction. © 2011 American Institute of Physics.


Collins H.,Carnegie Mellon University
Journal of High Energy Physics | Year: 2013

It is possible to define a general initial state for a quantum field by introducing a contribution to the action defined at an initial-time boundary. The propagator for this theory is composed of two parts, one associated with the free propagation of fields and another produced by the operators of this initial action. The derivation of this propagator is shown for the case of a translationally and rotationally invariant initial state. In addition to being able to treat more general states, these techniques can also be applied to effective field theories that start from an initial time. The eigenstates of a theory with interacting heavy and light fields are different from the eigenstates of the theory in the limit where the interactions vanish. Therefore, a product of states of the noninteracting heavy and light theories will usually contain excitations of the heavier state once the interactions are included. Such excitations appear as nonlocal effects in the effective theory, which are suppressed by powers of the mass of the heavy field. By appropriately choosing the initial action, these excitations can be excised from the state leaving just effects that would be produced by a local action of the lighter fields. © SISSA 2013.


Schaeffer J.,Carnegie Mellon University
Mathematical Methods in the Applied Sciences | Year: 2011

Batt showed that solutions of the Vlasov-Poisson system remain smooth as long as the particle speeds remain finite. Pfaffelmoser was the first to establish a bound on the particle speeds, completing the existence proof. Horst greatly improved this bound on the particle speeds. This article improves it further. © 2010 John Wiley & Sons, Ltd.


Grossmann I.E.,Carnegie Mellon University
Computers and Chemical Engineering | Year: 2012

Enterprise-wide Optimization (EWO) has become a major goal in the process industries due to the increasing pressures for remaining competitive in the global marketplace. EWO involves optimizing the supply, manufacturing and distribution activities of a company to reduce costs, inventories and environmental impact, and to maximize profits and responsiveness. Major operational items include planning, scheduling, real-time optimization and control. We provide an overview of EWO in terms of a mathematical programming framework. We first provide a brief overview of mathematical programming techniques (mixed-integer linear and nonlinear optimization methods), as well as decomposition methods, stochastic programming and modeling systems. We then address some of the major issues involved in the modeling and solution of these problems. Finally, based on the EWO program at the Center of Advanced Process Decision-making at Carnegie Mellon, we describe several applications to show the potential of this area. © 2012 Elsevier Ltd.


Massalski T.B.,Carnegie Mellon University
Materials Transactions | Year: 2010

Several interesting features in the study of stabilities of phases, and in phase transformations, are discussed. It is proposed that symmetry considerations related to the presence of magnetism in iron suggests that the respective phases, BCC alpha and FCC gamma, have in fact lower symmetries than cubic. A proposal is made that the symbol beta used in the past for the designation of the paramagnetic BCC iron should perhaps be returned as a feature in phase diagrams. The importance of the new concept of a 'pseudogap' in die electronic band structure, as a stabilizing electronic feature, is discussed in the light of the Hume-Rothery electron concentration rule. It is proposed that since the thermal activation is a major feature in the behavior of isothermal martensites, a more suitable designation for these types of phase transformations might be "thermally activated martensites'', or TAMs. Massive transformations are discussed briefly and it is emphasized that they present a specific example of an idiomorphic transformation process, not requiring the need for orientation relationships (ORs) between the parent and product phases. ©2010 The Japan Institute of Metals.


Minden J.S.,Carnegie Mellon University
Methods in Molecular Biology | Year: 2012

This chapter provides a brief historical perspective of the development of difference gel electrophoresis, from its inception to commercialization and beyond. © 2012 Springer Science+Business Media, LLC.


Dhumal N.R.,Carnegie Mellon University
Spectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy | Year: 2011

In the present work, we have studied the electronic structure, molecular electrostatic potential (MEP) and hydrogen bonding in DMSO-ethanol, DMSO-methanol and DMSO-water complexes by employing the MP2 method. Different conformers were simulated on the basis of possible binding sites guided by molecular electrostatic potential topology. The stronger hydrogen bonded interaction lowers the energy of the conformer. Molecular electron density topology and natural bond orbital analysis were used to explain the strength of interactions. Experimental vibrations are also compared with the calculated normal vibrations. Blue shift is predicted for SC vibration in experimental and theoretical spectra as well. Molecular electrostatic potential and topology are used to understand the interaction strength of the conformer. © 2011 Elsevier B.V. All rights reserved.


Saragih J.M.,CSIRO | Lucey S.,CSIRO | Cohn J.F.,Carnegie Mellon University
International Journal of Computer Vision | Year: 2011

Deformable model fitting has been actively pursued in the computer vision community for over a decade. As a result, numerous approaches have been proposed with varying degrees of success. A class of approaches that has shown substantial promise is one that makes independent predictions regarding locations of the model's landmarks, which are combined by enforcing a prior over their joint motion. A common theme in innovations to this approach is the replacement of the distribution of probable landmark locations, obtained from each local detector, with simpler parametric forms. In this work, a principled optimization strategy is proposed where nonparametric representations of these likelihoods are maximized within a hierarchy of smoothed estimates. The resulting update equations are reminiscent of mean-shift over the landmarks but with regularization imposed through a global prior over their joint motion. Extensions to handle partial occlusions and reduce computational complexity are also presented. Through numerical experiments, this approach is shown to outperform some common existing methods on the task of generic face fitting. © 2010 Springer Science+Business Media, LLC.


Thiessen E.D.,Carnegie Mellon University | Pavlik P.I.,University of Memphis
Cognitive Science | Year: 2013

Statistical learning refers to the ability to identify structure in the input based on its statistical properties. For many linguistic structures, the relevant statistical features are distributional: They are related to the frequency and variability of exemplars in the input. These distributional regularities have been suggested to play a role in many different aspects of language learning, including phonetic categories, using phonemic distinctions in word learning, and discovering non-adjacent relations. On the surface, these different aspects share few commonalities. Despite this, we demonstrate that the same computational framework can account for learning in all of these tasks. These results support two conclusions. The first is that much, and perhaps all, of distributional statistical learning can be explained by the same underlying set of processes. The second is that some aspects of language can be learned due to domain-general characteristics of memory. Copyright © 2012 Cognitive Science Society, Inc.


Li Y.-H.,Feng Chia University | Savvides M.,Carnegie Mellon University
IEEE Transactions on Pattern Analysis and Machine Intelligence | Year: 2013

Iris masks play an important role in iris recognition. They indicate which part of the iris texture map is useful and which part is occluded or contaminated by noisy image artifacts such as eyelashes, eyelids, eyeglasses frames, and specular reflections. The accuracy of the iris mask is extremely important. The performance of the iris recognition system will decrease dramatically when the iris mask is inaccurate, even when the best recognition algorithm is used. Traditionally, people used the rule-based algorithms to estimate iris masks from iris images. However, the accuracy of the iris masks generated this way is questionable. In this work, we propose to use Figueiredo and Jain's Gaussian Mixture Models (FJ-GMMs) to model the underlying probabilistic distributions of both valid and invalid regions on iris images. We also explored possible features and found that Gabor Filter Bank (GFB) provides the most discriminative information for our goal. Finally, we applied Simulated Annealing (SA) technique to optimize the parameters of GFB in order to achieve the best recognition rate. Experimental results show that the masks generated by the proposed algorithm increase the iris recognition rate on both ICE2 and UBIRIS dataset, verifying the effectiveness and importance of our proposed method for iris occlusion estimation. © 1979-2012 IEEE.


Hanneke S.,Carnegie Mellon University
Journal of Machine Learning Research | Year: 2012

We study the theoretical advantages of active learning over passive learning. Specifically, we prove that, in noise-free classifier learning for VC classes, any passive learning algorithm can be transformed into an active learning algorithm with asymptotically strictly superior label complexity for all nontrivial target functions and distributions. We further provide a general characterization of the magnitudes of these improvements in terms of a novel generalization of the disagreement coefficient. We also extend these results to active learning in the presence of label noise, and find that even under broad classes of noise distributions, we can typically guarantee strict improvements over the known results for passive learning. © 2012 Steve Hanneke.


Kadane J.B.,Carnegie Mellon University
Prevention Science | Year: 2015

Bayesian statistics represents a paradigm shift in statistical reasoning and an approach to analysis that is applicable to prevention trials with small samples. This paper introduces the reader to the philosophy behind Bayesian statistics. This introduction is followed by a review of some issues that arise in sampling statistics and how Bayesian methods address them. Finally, the article provides an extended illustration of the application of Bayesian statistics to data from a prevention trial that tested a family-focused intervention. © 2014, Society for Prevention Research.


Huh S.,Carnegie Mellon University
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention | Year: 2012

The detection of apoptosis, or programmed cell death, is important to understand the underlying mechanism of cell development. At present, apoptosis detection resorts to fluorescence or colorimetric assays, which may affect cell behavior and thus not allow long-term monitoring of intact cells. In this work, we present an image analysis method to detect apoptosis in time-lapse phase-contrast microscopy, which is nondestructive imaging. The method first detects candidates for apoptotic cells based on the optical principle of phase-contrast microscopy in connection with the properties of apoptotic cells. The temporal behavior of each candidate is then examined in its neighboring frames in order to determine if the candidate is indeed an apoptotic cell. When applied to three C2C12 myoblastic stem cell populations, which contain more than 1000 apoptosis, the method achieved around 90% accuracy in terms of average precision and recall.


Snider J.D.,Carnegie Mellon University
Magnetic resonance in chemistry : MRC | Year: 2012

A new strategy to assign diastereotopic protons was developed on the basis of residual dipolar couplings (RDCs) collected in compressed poly(methyl methacrylate) (PMMA) gels. A combination of 2D J-scaled BIRD HSQC and J-scaled BIRD HMQC/HSQC NMR experiments was used to collect the RDC data. In the proposed strategy, the first experiment is used to measure (1)D(CH) for methine groups, the sum of (1)D(CHa) + (1)D(CHb) for methylene groups and the average (1)D(CH3) value for methyl groups. In turn, the small molecule alignment tensor is calculated using these D values without the a priori assignment of CH(2) diastereotopic protons. The D values of each individual CH bond (CHa and CHb) of each methylene group in the molecule are then predicted using the calculated alignment tensor and these values compared with the results from the HMQC/HSQC experiment, leading to their unambiguous assignment. This strategy is demonstrated with the alkaloid strychnine that contains five methylene groups with diastereotopic protons, and our results fully agree with the previously reported assignment using combinations of permutated assignments. Copyright © 2012 John Wiley & Sons, Ltd.


Pinsky M.R.,University of Pittsburgh | Dubrawski A.,Carnegie Mellon University
American Journal of Respiratory and Critical Care Medicine | Year: 2014

It is often difficult to accurately predict when, why, and which patients develop shock, because signs of shock often occur late, once organ injury is already present. Three levels of aggregation of information can be used to aid the bedside clinician in this task: analysis of derived parameters of existing measured physiologic variables using simple bedside calculations (functional hemodynamic monitoring); prior physiologic data of similar subjects during periods of stability and disease to define quantitative metrics of level of severity; and libraries of responses across large and comprehensive collections of records of diverse subjects whose diagnosis, therapies, and course is already known to predict not only disease severity, but also the subsequent behavior of the subject if left untreated or treated with one of the many therapeutic options. The problem is in defining the minimal monitoring data set needed to initially identify those patients across all possible processes, and then specifically monitor their responses to targeted therapies known to improve outcome. To address these issues, multivariable models using machine learning data-driven classification techniques can be used to parsimoniously predict cardiorespiratory insufficiency. We briefly describe how these machine learning approaches are presently applied to address earlier identification of cardiorespiratory insufficiency and direct focused, patient-specific management. Copyright © 2014 by the American Thoracic Society.


Meiksin A.,University of Edinburgh | Whalen D.J.,Carnegie Mellon University
Monthly Notices of the Royal Astronomical Society | Year: 2013

Primordial stars are key to primeval structure formation as the first stellar components of primeval galaxies, the sources of cosmic chemical enrichment and likely cosmic reionization, and they possibly gave rise to the super-massive black holes residing at the centres of galaxies today. While the direct detection of individual Pop III stars will likelyremain beyond reach for decades to come, we show their supernova remnants may soon be detectable in the radio. We calculate radio synchrotron signatures between 0.5 and 35 GHz from hydrodynamical computations of the supernova remnants of Pop III stars in 107M⊙ minihaloes. We find that hypernovae yield the brightest systems, with observed radio fluxes as high as 1-10 μJy. Less energetic Type II supernovae yield remnants about a factor of 30 dimmer and pair-instability supernova remnants are dimmer by a factor of more than 10 000. Because of the high gas densities of the progenitor environments, synchrotron losses severely limit the maximum emission frequencies, producing a distinctive peaked radio spectrum distinguishable from normal galactic supernova remnant spectra. Hypernovae radio remnants should be detectable by existing radio facilities like eVLA and eMERLIN while Type II supernova remnants will require the Square Kilometre Array. The number counts of hypernova remnants at z > 20 with fluxes above 1 μJy are expected to be one per hundred square degree field, increasing to a few per square degree if they form down to z = 10. The detection of a z > 20 Type II supernova remnantbrighter than 1 nJy would require a 100-200 deg2 field, although only a 1-2 deg2 field for those forming down to z = 10.Hypernova and Type II supernova remnants are easily separated from one another by their light curves, which will enable future surveys to use them to constrain the initial mass function of Pop III stars. © 2013 The Authors. Published by Oxford University Press on behalf of the Royal Astronomical Society.


In this investigation we use THz time-scale spectroscopy to conduct an initial set of studies on myoglobin with the aim of providing further insight into the global, collective thermal fluctuations in the protein that have been hypothesized to play a prominent role in the dynamic formation of transient ligand channels as well as in shaping the molecular level basis for ligand discrimination. Using the two ligands O2 and CO, we have determined that the perturbation from the heme-ligand complex has a strong influence on the characteristics of the myoglobin collective dynamics that are excited upon binding. Further, the differences detected in the collective protein motions in Mb-O2 compared with those in Mb-CO appear to be intimately tied with the pathways of long-range allosteric communication in the protein, which ultimately determine the trajectories selected by the respective ligands on the path to and from the heme-binding cavity. © 2014 the Partner Organisations.


Hug G.,Carnegie Mellon University | Giampapa J.A.,U.S. Software Engineering Institute
IEEE Transactions on Smart Grid | Year: 2012

This paper introduces new analytical techniques for performing vulnerability analysis of state estimation when it is subject to a hidden false data injection cyber-attack on a power grid's SCADA system. Specifically, we consider ac state estimation and describe how the physical properties of the system can be used as an advantage in protecting the power system from such an attack. We present an algorithm based on graph theory which allows determining how many and which measurement signals an attacker will attack in order to minimize his efforts in keeping the attack hidden from bad data detection. This provides guidance on which measurements are vulnerable and need increased protection. Hence, this paper provides insights into the vulnerabilities but also the inherent strengths provided by ac state estimation and network topology features such as buses without power injections. © 2010-2012 IEEE.


Beladi H.,Deakin University | Rohrer G.S.,Carnegie Mellon University
Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science | Year: 2013

The grain boundary character distribution in a commercial IF steel has been measured as a function of lattice misorientation and boundary plane orientation. The grain boundary plane distribution revealed a relatively low anisotropy with a tendency for grain boundaries to terminate on low index planes having relatively low surface energy and large interplanar spacings. Although the most common grain boundary plane orientation was (111), grain boundaries terminated on higher index planes were sometimes found. For instance, at a misorientation angle of 60 deg about [111], symmetric {112} tilt boundaries were far more populous than [111] twist boundaries. The current observation revealed an inverse relationship between the measured populations and the previously reported grain boundary energies. © 2012 The Minerals, Metals & Materials Society and ASM International.


Avigad J.,Carnegie Mellon University | Harison J.,Intel Corporation
Communications of the ACM | Year: 2014

The article discusses how formal verification could become the new standard for rigor in mathematics with the help of computational proof assistants. Due to developments in computer science over the past few decades, it is now possible to achieve complete formalization in practice. Working with 'computational proof assistants,' users are able to verify substantial mathematical theorems, constructing formal axiomatic derivations of remarkable complexity. The notion of proof lies at the heart of mathematics. Although early records of measurement and numeric computation predate the ancient Greeks, mathematics proper is commonly seen as having begun with development of the deductive method, as exemplified by Euclid's Elements of Geometry.


Matyjaszewski K.,Carnegie Mellon University
Israel Journal of Chemistry | Year: 2012

Atom transfer radical polymerization (ATRP) is currently one of the most often used synthetic polymerization methods to prepare well-defined polymers with complex architecture. This review covers some fundamentals of copper-based ATRP, presents basic structure-reactivity correlation for initiators and catalyst complexes and discusses the radical nature of reactive intermediates. New ATRP initiating processes with ppm amounts of copper catalysts and various reducing agents are described together with recent electrochemically controlled ATRP and polymerization in aqueous homogeneous and dispersed media. Examples of polymers with precisely controlled architecture are presented together with the effect of variable amounts of catalysts on molecular weight distribution and morphology of nanostructured block copolymers. Some current and forthcoming applications of polymers made by ATRP are presented. Copyright © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.


Carver C.S.,University of Miami | Scheier M.F.,Carnegie Mellon University
Trends in Cognitive Sciences | Year: 2014

Optimism is a cognitive construct (expectancies regarding future outcomes) that also relates to motivation: optimistic people exert effort, whereas pessimistic people disengage from effort. Study of optimism began largely in health contexts, finding positive associations between optimism and markers of better psychological and physical health. Physical health effects likely occur through differences in both health-promoting behaviors and physiological concomitants of coping. Recently, the scientific study of optimism has extended to the realm of social relations: new evidence indicates that optimists have better social connections, partly because they work harder at them. In this review, we examine the myriad ways this trait can benefit an individual, and our current understanding of the biological basis of optimism. © 2014 Elsevier Ltd.


Laszlo S.,Carnegie Mellon University | Federmeier K.D.,University of Illinois at Urbana - Champaign
Psychophysiology | Year: 2011

Linking print with meaning tends to be divided into subprocesses, such as recognition of an input's lexical entry and subsequent access of semantics. However, recent results suggest that the set of semantic features activated by an input is broader than implied by a view wherein access serially follows recognition. EEG was collected from participants who viewed items varying in number and frequency of both orthographic neighbors and lexical associates. Regression analysis of single item ERPs replicated past findings, showing that N400 amplitudes are greater for items with more neighbors, and further revealed that N400 amplitudes increase for items with more lexical associates and with higher frequency neighbors or associates. Together, the data suggest that in the N400 time window semantic features of items broadly related to inputs are active, consistent with models in which semantic access takes place in parallel with stimulus recognition. Copyright © 2010 Society for Psychophysiological Research.


Griffiths R.B.,Carnegie Mellon University
Studies in History and Philosophy of Science Part B - Studies in History and Philosophy of Modern Physics | Year: 2013

It is shown that quantum mechanics is noncontextual if quantum properties are represented by subspaces of the quantum Hilbert space (as proposed by von Neumann) rather than by hidden variables. In particular, a measurement using an appropriately constructed apparatus can be shown to reveal the value of an observable A possessed by the measured system before the measurement took place, whatever other compatible ([. B, A] = 0) observable B may be measured at the same time. © 2013 Elsevier Ltd.


Guillen-Gosalbez G.,Rovira i Virgili University | Grossmann I.,Carnegie Mellon University
Computers and Chemical Engineering | Year: 2010

This paper addresses the optimal design and planning of sustainable chemical supply chains (SCs) in the presence of uncertainty in the damage model used to evaluate their environmental performance. The environmental damage is assessed through the Eco-indicator 99, which includes recent advances made in life cycle assessment (LCA). The overall problem is formulated as a bi-criterion stochastic non-convex mixed-integer nonlinear program (MINLP). The deterministic equivalent of such a model is obtained by reformulating the joint chance-constraint employed to calculate the environmental performance of the SC in the space of uncertain parameters. The resulting bi-criterion non-convex MINLP is solved by applying the epsilon constraint method. To guarantee the global optimality of the Pareto solutions found, we propose a novel spatial branch and bound method that exploits the specific structure of the problem. The capabilities of our modeling framework and the performance of the proposed solution strategy are illustrated through a case study. © 2009 Elsevier Ltd. All rights reserved.


Grossmann I.E.,Carnegie Mellon University | Guillen-Gosalbez G.,Rovira i Virgili University
Computers and Chemical Engineering | Year: 2010

Sustainability has recently emerged as a key issue in process systems engineering (PSE). Mathematical programming techniques offer a general modeling framework for including environmental concerns in the synthesis and planning of chemical processes. In this paper, we review major contributions in process synthesis and supply chain management, highlighting the main optimization approaches that are available, including the handling of uncertainty and the multi-objective optimization of economic and environmental objectives. Finally, we discuss challenges and opportunities identified in the area. © 2009 Elsevier Ltd.


Dinur I.,Weizmann Institute of Science | Guruswami V.,Carnegie Mellon University
Proceedings - Annual IEEE Symposium on Foundations of Computer Science, FOCS | Year: 2013

We develop new techniques to incorporate the recently proposed "short code" (a low-degree version of the long code) into the construction and analysis of PCPs in the classical "Label Cover + Fourier Analysis" framework. As a result, we obtain more size-efficient PCPs that yield improved hardness results for approximating CSPs and certain coloringtype problems. In particular, we show a hardness for a variant of hypergraph coloring (with hyperedges of size 6), with a gap between 2 and exp(2Ω( √ log logN)) number of colors where N is the number of vertices. This is the first hardness result to go beyond the O(logN) barrier for a coloring-type problem. Our hardness bound is a doubly exponential improvement over the previously known O(log logN)-coloring hardness for 2- colorable hypergraphs, and an exponential improvement over the (logN)Ω(1)-coloring hardness for O(1)-colorable hypergraphs. Stated in terms of "covering complexity," we show that for 6-ary Boolean CSPs, it is hard to decide if a given instance is perfectly satisfiable or if it requires more than 2Ω( √ log logN) assignments for covering all of the constraints. While our methods do not yield a result for conventional hypergraph coloring due to some technical reasons, we also prove hardness of (logN)Ω(1)-coloring 2-colorable 6-uniform hypergraphs (this result relies just on the long code). A key algebraic result driving our analysis concerns a very low-soundness error testing method for Reed-Muller codes. We prove that if a function β : Fm 2 → F2 is 2Ω(d) far in absolute distance from polynomials of degree m-d, then the probability that deg(βg) ≤ m-3d/4 for a random degree d/4 polynomial g is doubly exponentially small in d. Copyright © 2013 by The Institute of Electrical and Electronics Engineers, Inc.


Hitchens T.K.,Carnegie Mellon University
Magnetic resonance in medicine : official journal of the Society of Magnetic Resonance in Medicine / Society of Magnetic Resonance in Medicine | Year: 2011

Current diagnosis of organ rejection following transplantation relies on tissue biopsy, which is not ideal due to sampling limitations and risks associated with the invasive procedure.We have previously shown that cellular magnetic resonance imaging (MRI) of iron-oxide labeled immune-cell infiltration can provide a noninvasive measure of rejection status by detecting areas of hypointensity on T2*-weighted images. In this study, we tested the feasibility of using a fluorine-based cellular tracer agent to detect macrophage accumulation in rodent models of acute allograft rejection by fluorine-19 ((19) F) MRI and magnetic resonance spectroscopy. This study used two rat models of acute rejection, including abdominal heterotopic cardiac transplant and orthotopic kidney transplant models. Following in vivo labeling of monocytes and macrophages with a commercially available agent containing perfluoro-15-crown-5-ether, we observed (19) F-signal intensity in the organs experiencing rejection by (19) F MRI, and conventional (1) H MRI was used for anatomical context. Immunofluorescence and histology confirmed macrophage labeling. These results are consistent with our previous studies and show the complementary nature of the two cellular imaging techniques. With no background signal, (19) F MRI/magnetic resonance spectroscopy can provide unambiguous detection of fluorine labeled cells, and may be a useful technique for detecting and quantifying rejection grade in patients. Copyright © 2010 Wiley-Liss, Inc.


Rothstein I.Z.,Carnegie Mellon University
Nuclear Physics B | Year: 2012

In this paper we calculate the exact partition function for free bosons on the plane with lacunae. First the partition function for a plane with two spherical holes is calculated by matching exactly for the infinite set of Wilson coefficients in an effective world line theory and then performing the ensuing Gaussian integration. The partition is then re-calculated using conformal field theory techniques, and the equality of the two results is made manifest. It is then demonstrated that there is an exact correspondence between the Wilson coefficients (susceptibilities) in the effective field theory and the weights of the individual excitations of the closed string coherent state on the boundary. We calculate the partition function for the case of three holes where CFT techniques necessitate a closed form for the map from the corresponding closed string pants diagrams. Finally, it is shown that the Wilson coefficients for the case of quartic and higher order kernels, where standard CFT techniques are no longer applicable, can also be completely determined. These techniques can also be applied to the case of non-trivial central charges. © 2012 Elsevier B.V.


Griffiths R.B.,Carnegie Mellon University
Foundations of Physics | Year: 2012

Stapp's counterfactual argument for quantum nonlocality based upon a Hardy entangled state is shown to be flawed. While he has correctly analyzed a particular framework using the method of consistent histories, there are alternative frameworks which do not support his argument. The framework dependence of quantum counterfactual arguments, with analogs in classical counterfactuals, vitiates the claim that nonlocal (superluminal) influences exist in the quantum world. Instead it shows that counterfactual arguments are of limited use for analyzing these questions. © 2012 Springer Science+Business Media, LLC.


Leordeanu M.,Romanian Academy of Sciences | Sukthankar R.,Intel Corporation | Hebert M.,Carnegie Mellon University
International Journal of Computer Vision | Year: 2012

Graph matching is an essential problem in computer vision that has been successfully applied to 2D and 3D feature matching and object recognition. Despite its importance, little has been published on learning the parameters that control graph matching, even though learning has been shown to be vital for improving the matching rate. In this paper we show how to perform parameter learning in an unsupervised fashion, that is when no correct correspondences between graphs are given during training. Our experiments reveal that unsupervised learning compares favorably to the supervised case, both in terms of efficiency and quality, while avoiding the tedious manual labeling of ground truth correspondences. We verify experimentally that our learning method can improve the performance of several state-of-the art graph matching algorithms. We also show that a similar method can be successfully applied to parameter learning for graphical models and demonstrate its effectiveness empirically. © 2011 Springer Science+Business Media, LLC.


Guillen N.,University of Texas at Austin | Schwab R.W.,Carnegie Mellon University
Archive for Rational Mechanics and Analysis | Year: 2012

In this work we provide an Aleksandrov-Bakelman-Pucci type estimate for a certain class of fully nonlinear elliptic integro-differential equations, the proof of which relies on an appropriate generalization of the convex envelope to a nonlocal, fractional-order setting and on the use of Riesz potentials to interpret second derivatives as fractional order operators. This result applies to a family of equations involving some nondegenerate kernels and, as a consequence, provides some new regularity results for previously untreated equations. Furthermore, this result also gives a new comparison theorem for viscosity solutions of such equations which depends only on the L ∞ and L n norms of the right-hand side, in contrast to previous comparison results which utilize the continuity of the right-hand side for their conclusions. These results appear to be new, even for the linear case of the relevant equations. © 2012 Springer-Verlag.


Mankoff J.,Carnegie Mellon University
Interactions | Year: 2012

Sustainability in (Inter)Action provides a forum for innovative thought, design, and research in the area of interaction design and environmental sustainability. The forum explores how HCI can contribute to the complex and interdisciplinary efforts to address sustainability challenges. © 2012 ACM.


Settles B.,Carnegie Mellon University
Synthesis Lectures on Artificial Intelligence and Machine Learning | Year: 2012

The key idea behind active learning is that a machine learning algorithm can perform better with less training if it is allowed to choose the data from which it learns. An active learner may pose "queries," usually in the form of unlabeled data instances to be labeled by an "oracle" (e.g., a human annotator) that already understands the nature of the problem. This sort of approach is well-motivated in many modern machine learning and data mining applications, where unlabeled data may be abundant or easy to come by, but training labels are difficult, time-consuming, or expensive to obtain. This book is a general introduction to active learning. It outlines several scenarios in which queries might be formulated, and details many query selection algorithms which have been organized into four broad categories, or "query selection frameworks." We also touch on some of the theoretical foundations of active learning, and conclude with an overview of the strengths and weaknesses of these approaches in practice, including a summary of ongoing work to address these open challenges and opportunities. Copyright © 2012 by Morgan & Claypool.


Osaka I.,Hiroshima University | Takimiya K.,Hiroshima University | McCullough R.D.,Carnegie Mellon University
Advanced Materials | Year: 2010

New semiconducting copolymers based on benzobisthiazole show excellent environmental stability in high-humidity air, which is an unusual performance for semiconducting polymers, along with OFET mobilities of as high as 0.26 cm 2/Vs, even with disordered thin-film structures. With these unique features, these new copolymers are fascinating materials with high processability, mobility, and stability as active layers for printable electronics. Copyright © 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.


Wang C.,Carnegie Mellon University | Blei D.M.,Princeton University
Journal of Machine Learning Research | Year: 2013

Mean-field variational methods are widely used for approximate posterior inference in many probabilistic models. In a typical application, mean-field methods approximately compute the posterior with a coordinate-ascent optimization algorithm. When the model is conditionally conjugate, the coordinate updates are easily derived and in closed form. However, many models of interest-like the correlated topic model and Bayesian logistic regression-are nonconjugate. In these models, mean-field methods cannot be directly applied and practitioners have had to develop variational algorithms on a case-by-case basis. In this paper, we develop two generic methods for nonconjugate models, Laplace variational inference and delta method variational inference. Our methods have several advantages: they allow for easily derived variational algorithms with a wide class of nonconjugate models; they extend and unify some of the existing algorithms that have been derived for specific models; and they work well on real-world data sets. We studied our methods on the correlated topic model, Bayesian logistic regression, and hierarchical Bayesian logistic regression. Copyright © 2013 Chong Wang and David M. Blei.


Swendsen R.H.,Carnegie Mellon University
American Journal of Physics | Year: 2014

All teachers and students of physics have absorbed the doctrine that probability must be normalized. Nevertheless, there are problems for which the normalization factor only gets in the way. An important example of this counter-intuitive assertion is provided by the derivation of the thermodynamic entropy from the principles of statistical mechanics. Unnormalized probabilities provide a surprisingly effective teaching tool that can make it easier to explain to students the essential concept of entropy. The elimination of the normalization factor offers simpler equations for thermodynamic equilibrium in statistical mechanics, which then lead naturally to a new and simpler definition of the entropy in thermodynamics. Notably, this definition does not change the formal expression of the entropy based on composite systems that I have previously offered. My previous definition of entropy has been criticized by Dieks, based on what appears to be a misinterpretation. I believe that the new definition presented here has the advantage of greatly reducing the possibility of such a misunderstanding-either by students or by experts. © 2014 American Association of Physics Teachers.


Ramasubramaniam A.,University of Massachusetts Amherst | Naveh D.,Carnegie Mellon University
Physical Review B - Condensed Matter and Materials Physics | Year: 2011

Graphene islands with zigzag edges embedded in nitrogen-terminated vacancies in hexagonal boron nitride are shown to develop intrinsic magnetism and preferentially order antiferromagnetically. The magnetic moment of each graphene island is given by the numerical imbalance of carbon atoms on its two sublattices, which is in turn directly related to the size of the host defect. We propose a carrier-mediated model for antiferromagnetic coupling between islands and estimate Néel temperatures for these structures in excess of 100 K in some instances, with the possibility of attaining even higher temperatures at higher island densities. Our results suggest the possibility of designing molecular magnets via defect engineering of hexagonal boron nitride templates followed by trapping of carbon atoms in the defects. © 2011 American Physical Society.


Woods K.N.,Carnegie Mellon University
Physical Review E - Statistical, Nonlinear, and Soft Matter Physics | Year: 2010

THz spectroscopy is used to investigate the dynamics of the globular protein hen egg white lysozyme under varying hydration and temperature conditions. An analysis of the experimental spectra has revealed that the amount of solvent in the hydration shell has a strong influence on the low-frequency protein conformational dynamics and also the arrangement of hydrogen bonds in the protein secondary structure. Furthermore at a hydration level >0.2 we identify collective backbone fluctuations in the protein secondary structure that are not present at low hydration. It is possible that these solvent induced modes are important for the biological function of the protein. © 2010 The American Physical Society.


Platzer A.,Carnegie Mellon University
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

Logic is a powerful tool for analyzing and verifying systems, including programs, discrete systems, real-time systems, hybrid systems, and distributed systems. Some applications also have a stochastic behavior, however, either because of fundamental properties of nature, uncertain environments, or simplifications to overcome complexity. Discrete probabilistic systems have been studied using logic. But logic has been chronically underdeveloped in the context of stochastic hybrid systems, i.e., systems with interacting discrete, continuous, and stochastic dynamics. We aim at overcoming this deficiency and introduce a dynamic logic for stochastic hybrid systems. Our results indicate that logic is a promising tool for understanding stochastic hybrid systems and can help taming some of their complexity. We introduce a compositional model for stochastic hybrid systems. We prove adaptivity, càdlàg, and Markov time properties, and prove that the semantics of our logic is measurable. We present compositional proof rules, including rules for stochastic differential equations, and prove soundness. © 2011 Springer-Verlag Berlin Heidelberg.


Berger L.,Carnegie Mellon University
Physical Review B - Condensed Matter and Materials Physics | Year: 2011

The Elliott theory of spin relaxation in metals and semiconductors is extended to metallic ferromagnets. Our treatment is based on the two-current model of Fert, Campbell, and Jaoul. The d→s electron-scattering process involved in spin relaxation is the inverse of the s→d process responsible for the anisotropic magnetoresistance (AMR). As a result, spin-relaxation rate 1/τsr and AMR Δρ are given by similar formulas, and are in a constant ratio if scattering is by solute atoms. Our treatment applies to nickel- and cobalt-based alloys which do not have spin-up 3d states at the Fermi level. This category includes many of the technologically important magnetic materials. And we show how to modify the theory to apply it to bcc iron-based alloys. We also treat the case of Permalloy Ni80Fe20 at finite temperature or in thin-film form, where several kinds of scatterers exist. Predicted values of 1/τsr and Δρ are plotted versus resistivity of the sample. These predictions are compared to values of 1/τsr and Δρ derived from ferromagnetic-resonance and AMR experiments in Permalloy. © 2011 American Physical Society.


Li X.,Carnegie Mellon University
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems | Year: 2010

The aggressive scaling of integrated circuit technology results in high-dimensional, strongly-nonlinear performance variability that cannot be efficiently captured by traditional modeling techniques. In this paper, we adapt a novel L0-norm regularization method to address this modeling challenge. Our goal is to solve a large number of (e.g., 104 - 106) model coefficients from a small set of (e.g., 102 - 103) sampling points without over-fitting. This is facilitated by exploiting the underlying sparsity of model coefficients. Namely, although numerous basis functions are needed to span the high-dimensional, strongly-nonlinear variation space, only a few of them play an important role for a given performance of interest. An efficient orthogonal matching pursuit (OMP) algorithm is applied to automatically select these important basis functions based on a limited number of simulation samples. Several circuit examples designed in a commercial 65 nm process demonstrate that OMP achieves up to 25× speedup compared to the traditional least-squares fitting method. © 2006 IEEE.


This paper develops a gradient theory of single-crystal plasticity based on a system of microscopic force balances, one balance for each slip system, derived from the principle of virtual power, and a mechanical version of the second law that includes, via the microscopic forces, work performed during plastic flow. When combined with thermodynamically consistent constitutive relations the microscopic force balances become nonlocal flow rules for the individual slip systems in the form of partial differential equations requiring boundary conditions. Central ingredients in the theory are geometrically necessary edge and screw dislocations together with a free energy that accounts for work hardening through a dependence on the accumulation of geometrically necessary dislocations. © 2010 Elsevier Ltd. All rights reserved.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: SOFTWARE ENG & FORMAL METHODS | Award Amount: 149.02K | Year: 2013

As software pervades our society and lives, failures due to software
bugs become increasingly costly. Scalable approaches for
systematically checking software to find crucial bugs hold a key to
delivering higher quality software at a lower cost. Mera is a
methodology to scale model checking and symbolic execution which are two
powerful approaches for systematic software analysis and known to be
computationally expensive.

The project builds on two novel concepts: memoization, which allows
re-using computations performed across different checks to amortize
the cost of software analysis; and ranging, which allows distributing
the analysis into sub-problems of lesser complexity, which can be
solved separately and efficiently. Mera consists of three research
thrusts. First, the core memoization and ranging techniques for model
checking and symbolic execution are developed. Second, these
techniques are optimized in the context of different kinds of changes,
like the program code, expected properties, or analysis search-depth
parameters. Third, these techniques are adapted to effectively
utilize available resources for parallel computation using static and
dynamic strategies, such as work stealing. Mera will help improve
software quality and reliability thus holding the potential to provide
substantial economic benefits and to improve our quality of life.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: ROBUST INTELLIGENCE | Award Amount: 449.99K | Year: 2016

This project studies biologically-inspired architectures for visual recognition. The human visual system can perform a remarkable number of tasks, from estimating the 3D shape of an object that is grasped to inferring subtle differences between two similar makes and models of cars. Such diverse sets of visual tasks are required of a range of autonomous agents, including self-driving cars or humanoid robotics. Such autonomous platforms have the potential to increase general welfare and health of the overall population. This project attempts to build a computational model capable of such diverse visual tasks. Motivated by biological evidence, this project explores the use of feedback logic to enable such computational reasoning. The project provides research opportunities for both undergraduate and graduate students and for increasing diversity in the fields of computer and human vision.

This research focuses on development of a unified hierarchical probabilistic model that can be used to solve multiple fine-grained visual tasks. Feedforward hierarchical models, of which the most ubiquitous are Convolutional Neural Nets (CNNs), have demonstrated remarkable performance in recent history. This project introduces hierarchical models for vision-with-scrutiny tasks, such as 3D articulated pose estimation and part segmentation. Rather than focusing on increasing performance on established benchmark performance, this research provides a theoretical framework for analyzing bottom-up (feedforward) CNNs and imbuing them with novel top-down reasoning capabilities. It does so by exploring a link between three dominant but disparate paradigms for visual recognition: feedforward neural models, generative probabilistic models (Boltzmann machines), and discriminative latent-variable models (deformable part models). The models introduced in this proposal allow CNNs to be used for large-scale multi-task learning, where tasks span both coarse-grained tasks (such as rapid scene categorization) and fine-grained tasks (such as 3D articulated pose estimation). By addressing multiple fine-grained tasks with a single hierarchical architecture, resource requirements for memory and speed are vastly decreased, important for embedded visual perception applications such as autonomous robots and vehicles.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: Genetic Mechanisms | Award Amount: 170.00K | Year: 2016

This project focuses on telomerase, the enzyme that is responsible for making telomeres, structures that protect the ends of all eukaryotic chromosomes, reminiscent of the hard ends of shoelaces. The genetic mechanism of telomere replication by telomerase is better studied in yeast and mammalian cells; as far from each other as those two organisms are, they in fact represent relatively recent branches of eukaryotic phylogeny. In contrast, in this study an ancient, deep branching lineage of eukaryotes, kinetoplastid protists will be examined. The organism of choice, Trypanosoma brucei, has telomerase with unique structural and functional properties, and result should indicate how this enzyme works. The project will encourage full participation from women and underrepresented minorities in science. Apart from training graduate and postdoctoral trainees, this project will also implement a three-year undergraduate program (BIOKEYS) in both PIs laboratories; the major goal of this program will be to teach undergraduate researchers to value interdisciplinary sciences early in their research careers. In addition to gaining experience from designing experiments and problem solving using high-end technologies, students will be teaching and learning from each other as part of a research group.

Telomerase, a ribonucleoprotein enzyme, provides the major means for elongation of chromosome ends (telomeres), thus counteracting the loss of linear DNA ends in each cell cycle due to incomplete DNA replication by conventional DNA polymerases. Telomerase has two core components, the Telomerase Reverse Transcriptase (TERT) that catalyzes telomere elongation, and the telomerase RNA (TER), which provides a template for telomere DNA synthesis. The mechanisms of telomere elongation by telomerase are poorly understood in Trypanosoma brucei, a deep branching Kinetoplastid. Therefore, this project dissects structural, biochemical and genetic features of telomerase RNA in T. brucei to understand the mechanism of telomerase regulation in early eukaryotic species. The recent discovery of the T. brucei TER reveals novel features exclusive to deep branching eukaryotes, suggesting mechanistic differences in the process of telomere synthesis between T. brucei and higher eukaryotic organisms. Therefore, this project: (i) investigates telomerase RNA structure at a single nucleotide resolution using NMR and SHAPE chemistry, (ii) defines key TERT contact sites on TER that are essential for telomerase function in T. brucei using HITS-CLIP technology, and (iii) establishes the functional significance of TR domains by genetic manipulations and telomerase functional assays. Overall, this research will allow significant advances in understanding the mechanistic details of telomerase evolution in protists.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: I-Corps | Award Amount: 50.00K | Year: 2016

Lignin is a plant-derived biopolymer that is the main byproduct of paper and cellulosic ethanol production. While >100 million tons are generated annually, only a small fraction finds technological application. In plants, lignin acts as a binder between cellulose fibers, so it has natural interfacial functionality. The Washburn Lab at CMU has developed a strategy for preparing high-performance surfactants based on a lignin core grafted with water-soluble polymers. The I-Corps team will be focused on exploring applications of these lignin surfactants in delivering herbicides used in agriculture. Many herbicides have low solubility in water, and a broad range of chemicals is used to apply them in the field. However, many of these chemicals are toxic and persistent in the environment. In addition to being based on a renewable resource, lignin-based surfactants are biocompatible and biodegradable, making them a potentially high-impact chemical technology in agricultural applications.

This I-Corps team is exploring applications of this technology as an adjuvant for the delivery of agrochemicals. The combination of reducing surface tension, low solution viscosity, and strong affinity for aromatic compounds that are common herbicides make these an attractive candidate from a performance perspective, and the intrinsic natural biodegradability and lack of toxicity could address many of issues facing the current generation of surfactants used in agriculture. However, agrochemical adjuvants represent a complex class of chemicals, and it will be critical to interview experts in chemical companies that make other adjuvants, formulations companies that sell directly to farmers, and end-users to understand better their needs. In addition, preliminary pilot-scale manufacturing will be performed to prepare kilogram quantities of materials for field tests and to assess variables such as coupling efficiency and kraft lignin source on the properties of the final product.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: COMPUTATIONAL MATHEMATICS | Award Amount: 217.06K | Year: 2014

The very essence of science is to explain and understand natural phenomena in order to predict and forecast outcomes. The most successful predictions result when fundamental laws of nature are integrated into conceptual models of the phenomena of interest. Newtons development of mathematical tools to express many fundamental laws of nature has resulted in mathematical models with unparalleled predictive power. These models consist of complex systems of equations relating the physical quantities of interest and form the conceptual foundation of modern engineering and science. Solution of these complex systems of equations is a key technology needed to realize the potential of these theories, and the computational tools under investigation in this project are indispensable in this step of the modeling process. This project will enhance the computational tools used to simulate materials such as polymers, liquid crystals, and many biological components. Improved predictive capability of computational models will play an essential role in the development and manufacture of many next generation devices such as micro-mechanical devices, biological materials, and prosthetic organs. Predicting material response is essential to determine biological and/or physiological function, reliability, and durability of these devices. In addition to the technological developments, this project will also support the education and training of the next generation of scientists needed sustain the remarkable pace of discovery and our scientific leadership in these disciplines.

The focus of this proposal is the development and analysis of numerical schemes to simulate materials whose macroscopic response depends upon the state of their fine scale structure. This scenario is typical when material particles exhibit elasticity, attraction and/or repulsion, entropic interactions which can result in phase formation, and internal dissipation. At the macroscopic scale these effects are modeled with internal variables which couple to the dynamic equations of motion. This multi-scale character gives rise to many modeling, mathematical, and numerical challenges. Models of materials with microstructure involve formidable systems of partial differential equations which inherit the delicate balance between transport and inertial effects, configurational energy, and dissipation of the physical system. While the past two decades have witnessed the development of many algorithms and codes in the engineering and scientific computing communities to solve these equations, there are many gaps in the mathematical theory and very little analysis of their fundamental properties is available. In this situation is important to develop numerical schemes which faithfully inherit the complex interactions of the physical system. Experience has shown that this paradigm can lead to a deeper understanding of the current schemes and frequently leads to improved and simpler algorithms. This project will bring together tools from partial differential equations, continuum mechanics, and numerical analysis, to develop and analyze numerical schemes which simulate these systems.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: Mechanics of Materials and Str | Award Amount: 420.00K | Year: 2016

This award supports fundamental research to map nanocrystalline metal creep, for which there is limited understanding, as a function of temperature and stress. High strength metals are used to reduce weight in many structural applications including aircraft and automobile components. Typically, metals are polycrystalline materials with grain size on the order of one hundred micrometers. One of the most effective methods to increase metal strength is to reduce the grain size to the nanometer scale. Such nanocrystalline metals are extremely strong at higher temperatures but typically suffer from creep at these temperatures. Thermally stabilized nanocrystalline metals are strong candidates for improved protective coatings, electrical contacts and are of great interest in nuclear material applications. This project will be beneficial for the training of graduate and undergraduate students in mechanical property evaluation, nanofabrication, electrodeposition and electron microscopy. Outreach activities will also be performed to develop YouTube videos relating fictional super powers of comic book characters to real-life materials enhancements.

A micromachined creep test platform has been developed and will be optimized. Objective one is to synthesize the first creep maps of nanocrystalline metals over a wide range of temperature and stresses. Previous work suggests that there is a threshold stress for secondary creep that scales inversely with grain size. Objective two is to investigate if this threshold stress exists and to explore its dependence on grain size. Objective three is to determine the underlying diffusion and dislocation-based mechanisms, in part through the use of transmission electron microscopy. In summary, this project endeavors to significantly improve the fundamental understanding of creep in nanocrystalline metals.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: Secure &Trustworthy Cyberspace | Award Amount: 49.75K | Year: 2015

This joint workshop between US and Dutch researchers explores technical and societal topics in privacy. The two-day workshop is jointly organized with the sponsorship of the Netherlands Organization for Scientific Research (NWO) and NSF, and is scheduled for October 2 & 3 2015 in Washington DC.

The workshop will contribute to a thorough scientific understanding of the intercultural issues surrounding privacy which is necessary for societal debates between nations. The workshop participants will identify topics of mutual interest between US and Dutch researchers, for consideration as submissions to potential future funding opportunities.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: ECONOMICS | Award Amount: 547.82K | Year: 2014

Current issues related to the funding model for higher education bring to the forefront striking features of this market in the U.S.: the coexistence of competing public and private sectors; a high degree of differentiation in college characteristics within and across sectors; and latitude for charging differential tuition to students based on place of residence, household income, and academic qualifications. Surprising little research on the interaction between colleges, public and private, is available to inform the policy debate. This research provides a model that characterizes the interaction of providers of higher education and can be used to evaluate alternative public funding policies.
This research will advance the research frontier in, at least, three important dimensions. First, the PIs will provide a new framework that incorporates important features of competition between private and public universities. The framework will capture differences in objectives between public and private institutions, differences in resources, and differences in constraints on tuition and admission policies. Second, the PIs will undertake empirical testing of their model. They will assemble a new data set on public and private universities as well as on parental and student choices. They propose to develop new estimators that can deal with the complex pricing and sorting patterns generated by their model. Finally, they propose to explore the policy implications of their work. They will examine changes in state and federal financial aid policies, and their impact on student attainment and achievement, as well as access to higher education by disadvantaged groups of society.
The research develops, estimates, and applies a new framework for understanding the market for higher education and for studying education policy issues. This framework significantly advances the state-of-the art in research providing a detailed analysis of the differences in objectives and constraints faced by private and public universities. The proposed research will provide the first estimated model of public-private competition in higher education and is thus on the frontier of integrating theory, computation, and estimation. The general-equilibrium approach is particularly important for study of changes in state and federal policies and the associated changes in admission and tuition policies.
The research explores important current public policy issues. Research findings will be disseminated via conferences, seminars, and research papers to promote broad access to the results. Ph.D. students will participate in the research and develop related research, thus contributing both to their training and to the body of knowledge on the subject matter of this proposal.


Grant
Agency: National Science Foundation | Branch: | Program: STTR | Phase: Phase I | Award Amount: 224.85K | Year: 2014

This Small Business Technology Transfer Research (STTR) Phase I project will produce technology capable of constructing three-dimensional (3D) models of scenes accurately, rapidly, and at low cost. The technology leverages recent developments in computer vision and enhances them for operation under varying ambient conditions while increasing speed of model construction, thus enabling practical applications. Key system design requirements identified include: 1) ability to produce complete and accurate high fidelity models quickly; 2) robustness to varying ambient conditions; 3) ease of use by a non-specialist without extensive training; and 4) low cost. The proposed research objectives include: improvements to state-of-the-art algorithms for robust and efficient 3D modeling from images; the development of an approach for dense 3D reconstruction using different sensor modalities; the development of a sensor view point planning algorithm; the creation of techniques for the identification of missing model data; and the development of a system for merging data from different sources into a coherent 3D model of the environment. The work plan includes a thorough assessment of the system?s ability to reconstruct a scene accurately and completely, as well as the benefit of reconstructing scenes using sensors deployed by low-cost vehicles. The broader impact/commercial potential of this project will be the widespread use of 3D modeling for scene documentation. Systems currently available for this application are expensive, slow, bulky, and their use requires special training. The technology resulting from this work will enable the production and commercialization of systems that are faster, more affordable, and easier to use by non-experts, thereby simplifying their adoption by a larger number of law enforcers, prosecutors, insurers, and other government agencies. In the transportation domain, for example, a speedy scene reconstruction is essential to restore the flow of traffic when accidents happen. On busy highways, traffic backup can grow at a rate of up to a mile per minute of delay in clearing the accident site. This currently limits the complete documentation of accidents to only those cases where fatalities occur. The technology proposed will allow the documentation of larger number of incidents. Additionally, a great reduction in the time and complexity involved in determining the cause of accidents is expected. The potential long-term benefits of this reduction include a more rapid evolution of regulation and policy to deal with chronic causes of traffic accidents, which will expedite the implementation of road safety mechanisms.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: POLYMERS | Award Amount: 825.00K | Year: 2015

NON-TECHNICAL SUMMARY

This research project is focused on synthesis, characterization and potential applications of well-defined polymers containing bulky functional substituents (BFS). Until recently many polymers with precisely controlled architectures were prepared, yet essentially with small chemical substituents attached onto their molecules. The main aim of this project is to enlarge the range of these substituents to include various BFS such as inorganics, biomolecules or short polymer chains in the form of macromonomers. The resulting polymers will have a very high local concentration of functional groups and consequently could provide new materials with unusual thermal, mechanical, optoelectronic or biological properties. Unfortunately, direct polymerization of monomers with BFS is very challenging and therefore this research will explore new pathways to overcome these challenges. This project should impact not only science, but also education and our society. Undergraduate and graduate students, a postdoctoral fellow, and visitors from collaborative groups will work together in the PIs group and with industrial members of a consortium on specialized novel polymerization techniques. Some hybrid functional materials resulting from the proposed activities may be of commercial importance and should benefit our society. Dissemination of information will be accomplished by publications, presentations at national and international meetings and at the PIs web site.



TECHNICAL SUMMARY

This main aim of this research is synthesis, characterization and potential applications of polymers with bulky functional substituents (BFS) such as inorganic nanoparticles, short polymer chains, or biomolecules. BFS reduce thermodynamic and kinetic polymerizability of such monomers, and new pathways and reaction conditions to overcome these challenges will be explored. Three classes of monomers with BSF will be investigated: (i) inorganic systems with polyhedral oligomeric silsesquioxanes (POSS) with various substituents as well as gold nanoclusters with spacers linking them to polymerizable (meth)acrylate group; (ii) macromonomers based on poly(ethylene oxide) and poly(n-butyl acrylate) with (meth)acrylate function; (iii) biomolecules (proteins or oligonucleotides) linked to (meth)acrylamide polymerizable groups. They will be polymerized under the range of conditions (solvent, concentrations, pressure, chemical or physical activators, pre-assembly) to generate novel hybrid materials with high local concentration of bulky functional groups. Their properties (thermomechanical, catalytic, bio-relevant) will be thoroughly characterized and their potential applications will be explored. The generated information will be disseminated through the PIs website, timely publications, student presentations at national and international conferences, and also to industry through the industrial consortium at Carnegie Mellon. Some hybrid functional materials resulting from the proposed activities may be of commercial importance and should benefit our society.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: STATISTICS | Award Amount: 56.74K | Year: 2016

The constantly increasing dimensionality and complexity of modern data has motivated many new data analysis tools in various fields, and urgently call for rigorous theoretical investigation, such as robustness against different sources of model misspecification,uncertainty quantification in classification and prediction, and statistical performance guarantee of conventional methods under non-standard settings. Although most classical theory are not directly applicable to methods developed for complex data, partially due to highly specialized model assumptions and diversified algorithms, the profound statistical thinking carried in these long-established results can still provide deep theoretical insights. When combined with cutting-edge results in modern context such as random matrix theory, matrix concentration, and convex geometry, these classical theory will lead to novel principled methods for a general class of problems ranging from high dimensional regression and classification to network data analysis and subspace learning. All methods developed in the proposed research will be implemented as standard R packages freely available and will have high pedagogical value and will be used to develop new courses. The proposed research has applications in astronomy and medical screening data. The proposal also provides new inference tools for applied areas in genetics, psychiatry, brain sciences. Integrated educational activities include designing courses on new perspectives in nonparametric statistics and modern multivariate analysis.

The proposed work will further integrate classical nonparametric and multivariate analysis theory with modern elements in four major areas of statistical research, including assumption-free prediction bands in high dimensional regression; a generalized Neyman-Pearson framework for set-valued multi-class classification; statistical performance guarantee of some greedy algorithms in network community detection as well as goodness-of-fit tests for network model selection; and a unified singular value decomposition framework for structured subspace estimation formulated as a convex optimization problem. These research activities will lead to modernized nonparametric and multivariate analysis courses, featuring new theoretical frameworks such as computationally constrained minimax analysis, additional topics such as functional data analysis, and cutting-edge examples in genetics, brain imaging, traffic, and astronomy.


Grant
Agency: National Aeronautics and Space Administration | Branch: | Program: STTR | Phase: Phase I | Award Amount: 124.85K | Year: 2014

We propose to apply to planetary terrain mapping an alternative, multiresolution method, subdivision surfaces (subdivs), in place of conventional digital elevation maps (DEMs) and fixed-resolution meshes. The proposed research is innovative in that it presents a new setting for subdivs that demands novel extensions to subdivision algorithms, techniques and theory. The primary objectives of this work are to: (1) demonstrate suitability of subdivs as a representation for terrain data with highly varied spatial resolution and 3-D features; (2) demonstrate their ability to encode, and later re-register, local detail non-destructively through hierarchical edits; and (3) prototype a software user interface introducing new capabilities enabled by the subdiv representation. The expected benefits are: (a) higher-fidelity terrain visualization with reduced infrastructure requirements; (b) ability to visualize 3-D features, such as overhangs, missed in DEM's; (c) compact encoding with natural level-of-detail control for interactive viewing even on mobile devices; (d) greater algorithmic efficiency for non-visualization scientific computation; and (e) enablement of new software-tool capabilities for dynamic mapping of alternative local-terrain datasets, non-destructive experimentation, collaboration, and data traceability. The innovation also promises capability and reliability benefits to a surface robot by unifying terrain representations and enabling minimal upload of only incremental terrain details from the ground.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: COMM & INFORMATION FOUNDATIONS | Award Amount: 150.00K | Year: 2014

Datasets that are collected in engineering, social, commercial and other domains are becoming increasingly larger, complex and irregular in structure. There is an urgent need for the development of methods that formalize and automate analysis of such data and are capable of extracting valuable information. Recently, a theoretical framework called signal processing on graphs has emerged as a new approach to analyze data with irregular structure; it extends fundamental signal processing concepts to data residing on arbitrary graphs, as well as formulates data analysis problems as standard signal processing tasks. Moreover, as data often needs to be analyzed at multiple levels of detail, the investigators develop the fundamentals of the multiresolution analysis on graphs, extending the theory of discrete signal processing on graphs.

This research extends relevant signal processing concepts that are critical to the development of the multiresolution analysis, including signal translation, scaling, and sampling, to general graphs. The generalized definition is consistent with the classical multiresolution theory for time data and images; it also addresses additional challenges presented by structures and properties of graphs. Additionally, this research involves techniques and devices for the application and implementation of multiresolution analysis methods to data on graphs. The investigators also develop a set of general, theoretical methodologies that can later be instantiated and applied to datasets of different origin, nature and structure.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: PROBABILITY | Award Amount: 425.65K | Year: 2013

This project studies various phenomena related to diffusive tracer particles in a strong array of opposing vorticies. One particularly interesting aspect is a seemingly anomalous diffusive behaviour exhibited on intermediate time scales. On long time scales this anomalous diffusive behaviour should average to an effectively Brownian behaviour, recovering classical homogenization results. The project crucially uses tools from both Probability and PDE. The educational component of this project aims to further interaction between these two areas from the undergraduate to the post-graduate level.

Observing the movement of gravel particles shows intermittent periods of rest and travel. One aspect of this project studies an idealised system modelling this behaviour with the eventual aim of rigorously quantifying the rest and travel. This is closely related to questions about homogenization -- the approximation of inhomogeneous media with microscopic inhomogeneities by a homogeneous one. This project crucially uses tools from two connected but different areas in Mathematics: Probability and PDE. The educational component of this project aims to further interaction between these two areas from the undergraduate level all the way to the post-graduate level.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: Big Data Science &Engineering | Award Amount: 894.89K | Year: 2012

Tensors are multi-dimensional generalizations of matrices, and so can have non-numeric entries. Extremely large and sparse coupled tensors arise in numerous important applications that require the analysis of large, diverse, and partially related data. The effective analysis of coupled tensors requires the development of algorithms and associated software that can identify the core relations that exist among the different tensor modes, and scale to extremely large datasets. The objective of this project is to develop theory and algorithms for (coupled) sparse and low-rank tensor factorization, and associated scalable software toolkits to make such analysis possible. The research in the project is centered on three major thrusts. The first is designed to make novel theoretical contributions in the area of coupled tensor factorization, by developing multi-way compressed sensing methods for dimensionality reduction with perfect latent model reconstruction. Methods to handle missing values, noisy input, and coupled data will also be developed. The second thrust focuses on algorithms and scalability on modern architectures, which will enable the efficient analysis of coupled tensors with millions and billions of non-zero entries, using the map-reduce paradigm, as well as hybrid multicore architectures. An open-source coupled tensor factorization toolbox (HTF- Hybrid Tensor Factorization) will be developed that will provide robust and high-performance implementations of these algorithms. Finally, the third thrust focuses on evaluating and validating the effectiveness of these coupled factorization algorithms on a NeuroSemantics application whose goal is to understand how human brain activity correlates with text reading & understanding by analyzing fMRI and MEG brain image datasets obtained while reading various text passages.

Given triplets of facts (subject-verb-object), like (Washington is the capital of USA), can we find patterns, new objects, new verbs, anomalies? Can we correlate these with brain scans of people reading these words, to discover which parts of the brain get activated, say, by tool-like nouns (hammer), or action-like verbs (run)?
We propose a unified coupled tensor factorization framework to systematically mine such datasets. Unique challenges in these settings include
(a) tera- and peta-byte scaling issues,
(b) distributed fault-tolerant computation,
(c) large proportions of missing data, and
(d) insufficient theory and methods for big sparse tensors.
The Intellectual Merit of this effort is exactly the solution to the above four challenges.

The Broader Impact is the derivation of new scientific hypotheses on how the brain works and how it processes language (from the never-ending language learning (NELL) and NeuroSemantics projects) and the development of scalable open source software for coupled tensor factorization. Our tensor analysis methods can also be used in many other settings, including recommendation systems and computer-network intrusion/anomaly detection.

KEYWORDS:
Data mining; map/reduce; read-the-web; neuro-semantics; tensors.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 418.59K | Year: 2014

This effort seeks to understand and develop strategies for effective movement in biological and synthetic locomoting systems. Gaits are a fundamental aspect of animal locomotion; examples include a horses walking, a fishs strokes, and a snakes slithering. In these motions, the animals undergo cyclic motions which interact with the surrounding environment to gain a net displacement over each cycle. The efficacy of such gaits suggests they form a core capability in locomotion of mechanical systems. Understanding the principles of gait-based locomotion offers two opportunities: to gain deep insight into biological processes and to create sophisticated synthetic locomotors to send mechanical systems into dangerous and dirty environments. To gain this insight, questions arise: how to model locomotion, and with this model, how to both evaluate and design gaits to achieve desired locomotive capabilities? In this project, the focus will be on limbless locomotors, including snakes, slender lizards, bacteria, spermatozoa and nematode worms. Limbless locomotor controllers for confined space applications, such as search and rescue in collapsed buildings and landslide debris, will be developed.

The investigators preliminary work reveals that geometric mechanics allows intuitive understanding of how and why gaits, produce successful locomotion. Much of the prior work with geometric tools, however, provided computationally burdensome approaches to design gaits: choose parameterized basis functions for gaits, simulate the motion of the system and then optimize the input parameters to find gaits that meet the design requirements. Such optimization with forward simulation is computationally expensive. Moreover, existing geometric approaches ignore real world considerations such as body-shape and granular (e.g., dirt) interaction between the mechanism and the environment. Therefore, the intellectual merit of this work is to advance the design and evaluation of gaits for complex systems by representing complex shapes as a basis of curvature functions, while all along empirically deriving from biological observation linear relationships between these parameters and the resulting displacement in granular media. Calculations will then take minutes rather than the days needed for multi-particle discrete element method (DEM) simulation, mitigating the challenges inherent in performing many experiments on real mechanical systems. This work will contribute to a new understanding of biological locomotors as well as help create life-life locomotion in mechanical systems.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 463.16K | Year: 2014

How do our brains take the light entering our eyes and turn it into our experience of the world around us? Critically, this experience seems to involve a visual vocabulary that allows us to understand new scenes based on our prior knowledge. The investigators explore the nature of this visual language, exploring the specific computations that are realized in the brain mechanisms used for scene perception. The work combines data from state-of-the-art computer vision systems with human neuroimaging to both predict brain responses when viewing complex, real-world scenes, and to analyze and understand the hidden structure embedded in real-world images. This effort is essential for building a theory of how we are able to see and for improving machine vision systems. More broadly, biologically-inspired models of vision are essential for the effective deployment of intelligent technology in navigation systems, assistive devices, security verification, and visual information retrieval.

The artificial vision system adopted in this research is highly data-driven in that it is learning about the visual world by continuously looking at real-world images on the World Wide Web. The model, known as NEIL (Never Ending Image Learner, http://www.neil-kb.com/), leverages cutting-edge big-data methods to extract a vocabulary of scene parts and relationships from hundreds of thousands of images. The relevance of this vocabulary to human vision will then be tested using both functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG) neuroimaging. The hypothesis is that the application of prior knowledge about scenes expresses itself through learned associations between the specific parts and relations forming the vocabulary for scene perception. Moreover, different kinds of associations may be instantiated within distinct components of the functional brain network responsible for scene perception. Overall, this research will build on a recent, highly-successful artificial vision system in order to provide a more well-specified theory of the parts and relations underlying human scene perception. At the same time, the research will provide information about the human functional relevance of computationally-derived scene parts and relations, thereby helping to refine and improve artificial vision systems.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 88.44K | Year: 2014

New interest in spoken dialog research has recently been generated by the advent of SIRI(TM). This application has ignited the imagination of many who have begun to believe that speaking to automatic assistants is possible and useful. In the near future everyone will need to use speech to communicate with technology such as Google Glass and smart watches which cant use a keyboard. Using speech will also give wider access to technology, for example to the elderly and the disabled. It will serve as an economic motivator, spurring new technology and technology-based products.
The vision of the Real User Speech Data project and community infrastructure (RUSD) is that it will serve the spoken dialog community by sparking the creation of streams of spoken dialog data from real users, by distributing data and by helping researchers access spoken dialog platforms where they can compare their findings with those of others. It will link high school and undergraduate students (who have lived with the technology all of their lives, and can imagine how they would like to talk to technology), with technology experts who can implement their ideas. RUSD will serve the needs of the spoken dialog community at the same time by helping this community to use these novel and useful speech applications to record and collect speech of their users. RUSD will help to distribute this speech data to the community. It is essential in making successful systems as it will be used to retrain them (adding data makes the statistical representations more precise) and to assess them (comparing systems helps find which novel techniques improve the systems) and to create new ones. In this manner, RUSDs contribution will lead to both higher quality research and to widespread use of speech technology. The current planning project is dedicated to the discussion of the future RUSD with the research community which includes a hands-on Challenge contest to finalize RUSDs functional requirements, and to the preparation of a comprehensive RUSD proposal to the CISE Research Infrastructure Program.


The overarching goal for the RUSD project and community infrastructure is sparking new initiative that will generate data and research platforms for the community. It will relate high school and undergraduate students ideas for novel, transformational applications of speech technology with the current spoken dialog communitys technology. The resulting real applications will provide the streams of data from real users who are interested in repeatedly using an application because it helps them in some way (solving some task, or entertaining them). This will allow researchers to make their systems more robust.
Unsolved basic research issues in spoken dialog include: signal processing in noise, recognition of groups of difficult users (like the elderly and non-native speakers), management of complex dialogs (i.e. in meetings and with agents), and the automatic use of meta-linguistic information such as prosody. RUSD will help support new applications so that they can be used as research platforms where these and other issues can be explored. The Dialog Research Center (DialRC) has provided this service for one application, Lets Go, which gives bus information in Pittsburgh. Since then research has evolved, necessitating real dialog systems in more challenging areas. One site cannot create and run all of them. RUSD can help the community to create these systems, insuring that they meet the needs of the community and obtaining consensus on a unified infrastructure.
The planning phase of RUSD includes: (1) querying the spoken dialog community to define its needs going forward; (2) conducting a Challenge to determine interest in producing real streams of data and platforms; (3) gathering tools that can be used for the Challenge systems and ensure their documentation; (4) preparing the full proposal to the CISE Research Infrastructure Program for RUSDs design and implementation.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 58.36K | Year: 2015

1541807
Lowry

With ever increasing global demands for food, water, and energy, there is a critical need in the U.S. to identify sustainable solutions for simultaneously achieving energy, water, and food security. Agriculture accounts for approximately 70 % of all freshwater use, and 10 % of energy consumption, globally. Therefore, the agroecosystem lies at the heart of the energy-water-food nexus, and systems-level improvements in performance offers one of the greatest opportunities towards energy, water, and food security. Recent advances in nanoscale science and engineering offer unprecedented opportunity to reduce the energy and water inputs for food production, and to provide cost-effective, water and energy conservative technologies for reducing the environmental footprint of agriculture.

The PIs propose to convene a multidisciplinary group of faculty, students, and researchers from the USDA and the nanotechnology and agrochemical industries in a 2-day workshop in Pittsburgh, PA to identify the most promising groundbreaking opportunities for nanotechnology to increase sustainability at the food-water-energy nexus, and to identify the most pressing scientific, engineering, and social challenges that must be overcome to realize those benefits. Nanotechnology now offers unprecedented capabilities to clean and recycle water and wastewater, harness solar energy, create novel miniaturized chemical sensors and integrated sensor networks, improve food preservation, and provide targeted delivery of pesticides and nutrients. These advances can provide the tools needed for data-driven precision agriculture, can substantially improve resource utilization, and can lower the environmental footprint of food production. However, realizing this goal will require improved understanding of the complex relationships between food, water, energy, and society. This pioneering workshop will bring together leading scientists and engineers from a range of research fields comprising the food-energy-water nexus to identify the most promising opportunities for nanotechnologies to improve overall agroecosystem performance, and to identify the scientific and engineering challenges currently inhibiting widespread applications of nanotechnology in food production. The PIs will apply the pseudo-nominal group technique during the workshop to develop a prioritized list of the most promising emerging opportunities for nanotechnology at the food-energy-water nexus, and to elaborate on how these opportunities should direct research efforts toward the most critical research needs, and potential impacts, in the next five to ten years. This workshop experience will foster dialogue among diverse participants around the most pressing scientific obstacles to implementing nanotechnology for sustainable solutions at the food-energy-water nexus, while also drawing on understanding of the myriad economic, infrastructural, and social constraints that also define the boundaries of both challenges and potential opportunities for within the food-energy-water nexus


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: SOFTWARE & HARDWARE FOUNDATION | Award Amount: 499.95K | Year: 2016

Software systems are becoming more ubiquitous and critical to the functioning of our lives. An increasingly important requirement is to maintain high availability of these systems even in the face of changing requirements, faults, and resources. To address that concern, system developers today incorporate hand-written run-time adaptation strategies to automatically keep a system functioning effectively. However, as software systems grow in both complexity and ubiquity, and as the rate of technological change continues to increase, manual approaches cannot keep up. We must instead treat the evolution of adaptation strategies as a first-order concern. This research develops new mechanisms to automatically adapt and evolve the adaptation strategies themselves. Our high-level approach is to reuse previous domain or expert knowledge to inform the construction of flexible strategies, able to adapt to unanticipated changes and to various potential dimensions of system or environmental change.

Future-generation software systems will need to automatically optimize for multiple interacting, difficult-to-measure, and evolving qualities, properties, and priorities. Existing work provides methods for constructing complex software systems that can adapt to the changing of certain circumstances such as changing environmental conditions, infrastructure availability, or user demands,
while continuing to provide service at required quality levels. Our motivating insight is that stochastic search methods are especially promising for
self-adaptive software systems, and in particular for tackling the evolution of self-adaptation strategies, as evidenced in part by recent work that scales such techniques to complex source-level software problems. This research develops a principled foundation for the evolution of adaptation strategies in the self-adaptive domain, using stochastic search. The resulting family of techniques reuses, recombines, and otherwise builds upon previous knowledge about a given system to adapt to four major potential change dimensions: (1) the systems architecture and deployment; (2) the tactics that can be deployed in an adaptation scenario, including mechanisms to choose between them and information regarding their applicability, costs, effects, success likelihood, etc.; (3) the systems quality goals, and their relative priorities; and (4) the environmental assumptions that control the context in which the system is deployed. The unifying factor in each of these strategies is the existence of previous domain or expert knowledge that can be leveraged for evolving adaptive strategies moving forward.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: Materials Eng. & Processing | Award Amount: 475.81K | Year: 2014

Steel is a metallic material with iron and carbon as the primary components. The properties of steel can vary greatly depending on carbon level and processing. Steels with ultra-high carbon levels behave more like ceramics than metals; they are extremely hard and very brittle. These steels are primarily utilized in tool components to shape softer steels and other metals, but could see greater usage if their brittleness can be reduced. To date, brittleness has been reduced by cycles of heating and forming, methods that are challenging to implement on a large production scale. It is possible to decrease brittleness in these materials by simply heating alone; however existing heat treatments are energy intensive and far from optimized. This award supports fundamental research on the structural transformation behavior of ultra-high carbon steels during heating. The research utilizes novel experimental methods to observe changes to the steel structure in real time during heating and to develop new metrics that can be used to optimize processing and properties. These fundamental studies will lead to improved heat treatment strategies that will enable production of ultra-high carbon steels with improved combinations of properties while also reducing energy usage. Results from this research will also introduce materials science undergraduate students to fundamental topics of modern steelmaking and an outreach component will expose middle and high school students to digital image analysis and methods used to characterize materials.

Ultra-high carbon steels can exhibit superior hardness and strength compared to other steels of lower carbon contents. However, the high fraction of carbides in these steels render them too brittle for many structural applications. This project seeks to investigate microstructural transformations at elevated temperatures in steels containing ultrahigh carbon contents (1.7-2.3 weight percent) and the impact of these transformations on mechanical properties: toughness, hardness, strength and ductility. This study will apply in-situ confocal scanning laser microscopy and ex-situ electron microscopy to study and quantify microstructural evolution, specifically examining the effects of time, temperature and alloy chemistry on carbide dissolution kinetics and their final distribution in heat treated steels. A comprehensive assessment of important mechanical properties will be correlated to the final microstructural state. These pursuits will result in an analytical model that can accurately describe carbide dissolution and precipitation behavior at temperature ranges employed for heat treatments of ultra-high carbon steels.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 50.00K | Year: 2014

Getting robots to perform useful work in uncertain environments is still a grand challenge in robotics. One of the biggest possible impacts of this ability would be to enable the development of new co-robots that are nurses, handymen, and butlers for the elderly and infirm. Other important applications would be co-robots that can help find and evacuate the injured during first responses to disasters, improvements in design and control of active prosthetic devices and exoskeletons, and locomotion and manipulation behaviors for collaborative human-robot work. This proposal requests funds to organize a workshop focused on developing new planning and control methods to extend robots? abilities to transport themselves to work sites and then to perform useful physical work.

A two-day workshop is proposed to encourage collaboration between the research communities of robotic manipulation and locomotion, both in academia and industry. The workshop is motivated by the fact that, although the underlying physical principles driving locomotion and manipulation are similar, the techniques developed by the two communities are not. The workshop will analyze the reasons for these differences and explore ideas to bring them closer, with the goal of cross-pollinating advances in both fields. In addition, the workshop will consider the impact of industrial efforts to address these problems.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: CROSS-EF ACTIVITIES | Award Amount: 42.70K | Year: 2015

The study of genomes is a critical and rapidly growing component in understanding the variability of life, biological functions, population dynamics, and how organisms respond to external influences. Genomics has qualitatively improved our ability to investigate biological dynamics and to make important discoveries that are the foundations for understanding topics such as environmental change, developing and protecting crops, and improving health outcomes. Genome analysis, however, is a significant challenge for the practicing biologist. Most biologists who need to undertake genome science are not sufficiently expert in the relevant analytical tools, or understand the complex workflows required to get from the initial data generated by sequencers to a biologically meaningful analyzed result. In addition, few have access to the supercomputing resources and large-scale storage required for processing and managing genomics data. The National Center for Genome Analysis Support (NCGAS) addresses these challenges by providing an integrated service comprised of expert consulting and educational services, hardened and optimized software available through easy to use web-based workflow management tools, large memory supercomputers, and large scale data storage and publishing facilities. These resources are particularly useful for researchers from smaller, and minority serving, institutions that typically do not have access to the required expertise and cyberinfrastructure, yet whose investigations are equally important. Since its inception in 2011, the NCGAS has supported over 80 research projects representing over $61M in funded research. It engaged in 51 training events that served 691 individuals, of which 241 were from traditionally underserved populations.

The NCGAS (http://ncgas.org) was established in 2011 through a National Science Foundation ABI development award to help the national research community complete genomics research that requires data management and computational infrastructures at scale. NCGAS is a partnership among the Indiana University Pervasive Technology Institute, the Pittsburgh Supercomputing Center, the Texas Advanced Computing Center, and the San Diego Supercomputing Center. It meets the technology challenges of modern genome science by providing excellent bioinformatics consulting services for genome analysis, particularly genome and transcriptome assembly, including research design, data analysis and visualization. It optimizes, supports, and delivers genome analysis software on national supercomputing systems such as those funded by the NSF eXtreme Digital (XD) program and coordinated by the eXtreme Science and Engineering Discovery Environment (XSEDE) and the Open Science Grid (OSG). The NCGAS maintains and supports easy-to-use gateways, including Galaxy web portals, for genome analysis workflows that lower barriers for scientist to create, execute, document, and share genomics analyses. It distributes software tools for genome analysis to research computing facilities and the general research community so that IT managers can more easily install these tools on their systems. It provides long-term archival storage services. The NCGAS provides a digital library resource for the dissemination of data sets, publications, reports, or collections of files that will allow research to be visible and data to be re-used for decades to come. It delivers education and outreach programs on genome analysis, interpretation, and data management to biology faculty and students nationally. These programs will enhance the technology literacy of practicing scientists and help grow the bioinformatics workforce. These services are particularly available to smaller institutions across the country without access to supercomputers, bioinformatics expertise, or training. The NCGAS will enable breakthroughs that would not be possible without advanced cyberinfrastructure support.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: PROCESS & REACTION ENGINEERING | Award Amount: 310.22K | Year: 2015

Gounaris, 1510787

This project addresses the mitigation of risk in the context of Process Scheduling Optimization (PSO) in view of uncertainty in problem parameters. The term PSO refers to a family of decision-making problems that are prevalent in the chemical process industries and which are typically embedded in the Manufacturing Execution System supervising a plants operations. The archetypal setting is the one where a set of limited available resources (e.g., equipment, personnel, raw materials, utilities) needs to be coordinated (scheduled) along a time horizon so as to meet a number of production goals. Identifying optimal solutions, and sometimes even obtaining a single feasible solution, of a PSO instance is generally a challenging task. This is due to the various compounding combinatorial complexities involved, including complexities stemming from the plants topology (flowsheet), complicated production recipes, or other operational restrictions (market-related, regulatory, etc.). The objective in PSO is typically the maximization of profit (produce as much as you can within a limited amount of time) or the minimization of makespan (produce a fixed amount as soon as possible), though additional objectives, such as the balancing of resource utilization load or the minimization of environmental footprint, can also be considered.

The technical objective is to develop an Adjustable Robust Optimization (ARO) framework for the systematic treatment of uncertainty in PSO. PSO involves the coordination of limited available resources along a time horizon so as to meet a number of production goals. Identifying optimal solutions of a PSO instance is generally a challenging task, further complicated by the fact that it is of practical interest that such production management systems take into account uncertainties in input data, since failure to do so may lead to solutions that are infeasible or highly suboptimal. This project applies ARO, a risk mitigation methodology extending the paradigm of Robust Optimization (RO) that seeks to optimize the problem in view of a worst-case scenario, as dictated by an uncertainty set. But unlike RO, which results in a static, here-and-now solution that is often overly conservative, ARO results in a more flexible--and generally more profitable--solution policy by adjusting the decisions on the actual realizations of the uncertain parameters that have already occurred and been observed by the time of the decision.

Effective algorithms to mitigate technical and financial risk in the process industries can play an important role in the competitiveness, product quality and sustainability of the U.S. manufacturing base. Exploiting efficiencies in process operations limits environmental impact as well as promotes occupational health and safety. Adopting these innovations could provide tangible benefits to individual companies by materializing efficiencies in their utilization of process equipment, raw materials and personnel. This could be particularly useful for small companies, which cannot readily develop an in-house framework suitable to their setting. There is also the potential to enhance products of software vendors in the sector of manufacturing and enterprise resource planning. Potential educational benefits will be in generating material for a relevant course and creating an educationally-focused PSO-themed software applet. All students will receive training in production management, optimization methods and algorithms, uncertainty quantification and analysis, and scientific computation.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: RES IN NETWORKING TECH & SYS | Award Amount: 109.77K | Year: 2016

The security, performance, and availability of our critical network infrastructures relies on the correct implementation of different policy goals. Network operators realize these goals by composing and configuring diverse network appliances such as routers, firewalls, intrusion prevention systems, and web proxies. Unfortunately, this process of managing networks is extremely challenging, error-prone, and entails significant manual effort and operational costs. Configuration and implementation errors could have significant consequences as it can degrade network performance, induce downtime for critical infrastructures, and cause violations of key security postures. Systematically identifying and diagnosing potential violations has been, and continues to be, a fundamental challenge. This project will develop a principled framework to check if a network setup correctly implements a given suite of policies and to help operators proactively and automatically diagnose and localize the sources of policy violations.

Checking policy violations is hard even for simple reachability properties (e.g., can A talk to B) in todays networks. Furthermore, next-generation technologies such as software-defined networking and network functions virtualization are poised to enable richer dynamic policies (e.g., if a host generates too many connections, subject it to deeper inspection) and also introduce new sources of complexity (e.g., elastic scaling, software bugs). Existing approaches in network testing and verification have fundamental expressiveness and scalability challenges in tackling dynamic policies and stateful elements. To address these challenges, the research will include developing a model-based testing framework that will lead to fundamental advances in network semantics, modeling, testing, and diagnosis. To this end, the project will design: (1) new formal semantics to express dynamic policies over stateful networks; (2) expressive-yet-efficient abstractions for modeling the behavior of advanced network functions; (3) techniques to synthesize models of network functions; (4) scalable symbolic execution algorithms to generate test cases; and (5) efficient diagnosis algorithms to validate tests and localize violations.

Broader Impacts: The proposed research and education activities will develop testing tools and abstractions, which is a new capability that bolsters networking education, research, and practice. The project will inspire future educators and practitioners to apply systematic network testing as an integral part of their workflow to guarantee the security, performance, and availability our critical infrastructures. The project will develop new educational modules that will be integrated into the research with undergraduate and graduate-level classes and undergraduate capstone classes. The research and education efforts will engage undergraduates and underrepresented communities. The project will create unique opportunities for interdisciplinary training of the future workforce across the domains of networking, security, programming languages, and formal verification. The project will also design novel security games and education modules for K-12 students and teachers to highlight the importance of network testing for securing our critical infrastructures. Finally, the project deliverables (code, models, and education modules) will be released under open-source licenses as enablers for other researchers and educators, and to help transition the ideas from research to practice.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: COGNEURO | Award Amount: 246.71K | Year: 2014

How does someone learn a complex skill that unfolds over time, such as learning to play a piano sonata? This ability entails interacting processing levels, including conceptual knowledge (e.g., the notes of the melody on the sheet of music) and motor production (e.g., the actions of physically pressing the piano keys). This research program will combine computational models, neuroimaging, and brain stimulation methods to explore how these two levels of sequential skill knowledge are acquired by interacting brain systems. The work takes advantage of the fact that motor learning leaves a signature in the timing movements, called chunking.

The impact of this project extends from clinical rehabilitation to basic models of brain function. A hallmark symptom of some neurodegenerative conditions, like Parkinsons disease, is a difficulty in learning new skills. Understanding how skill learning occurs in the healthy brain can provide critical insights into how it is affected in neurological conditions. Scientifically, this research program will also attempt to bridge two largely independent literatures in cognitive science (sequential skill learning) and neuroscience (basal ganglia plasticity), providing a biologically meaningful foundation for well established psychological phenomena. Finally, by producing new tools and novel data sets that will be made publicly available, the work will integrate with the broader open-science community that seeks to foster the scientific enterprise by improving access to tools and data.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: Theory, Models, Comput. Method | Award Amount: 480.00K | Year: 2015

Markus Deserno from Carnegie Mellon University is supported by the Chemical Theory, Models and Computational Methods Program of the Chemistry Division (CHE) and the Condensed Matter and Materials Theory Program of the Division of Materials Research (DMR) to develop computational approaches to predict the elastic properties of lipid membranes and to understand how protein filaments adsorbed onto these membranes exert forces upon them. These membranes, a combination of fat-like molecules, cholesterol and proteins, are complex entities across which metabolic intermediates enter and exit cells. Most cellular membranes require frequent changes of their shape or even connectivity in order to execute their wide spectrum of biological functions. The underlying mechanics depends on a small number of parameters, which fully characterize both the energetic requirements for such deformations as well as the forces transmitted by them. Deserno and his research group develop a set of new simulation strategies for predicting these parameters in computer simulations. The aim is to gain access to physical parameters that have been controversial or notoriously hard to obtain. They also develop quantitative approaches to investigate one of the most common processes by which membrane connectivity is changed. Graduate and undergraduate students working on these projects will gain experience in soft matter, biophysics, simulation, and continuum modeling. Ongoing outreach projects with both middle- and high schools in Pittsburgh will be extended by developing a lecture series closely tied with hands-on experimentation that develops the concept of elastic sheets and beams.

This award supports theoretical and computational research and education to (1) measure emergent continuum-elastic properties of lipid membranes from molecular-level simulations via consistent and model-free scale bridging and to (2) develop theoretical techniques for describing the interaction of semi-flexible polymers with curved surfaces. In (1) the research involves expanding the information obtainable from simulating buckled membranes to extract not only the bending modulus, but identify its entropic contribution, the position of the pivotal plane of a single leaflet, and the magnitude of the spontaneous monolayer curvature. These techniques are applied to computational membrane models spanning a wide range of resolution, from atomistic to highly coarse-grained. Particular applications include lipid bilayers that are strongly stiffened when entering a gel phase, or strongly softened by trace amounts of small peptides. Deserno and coworkers also aim to measure the Gaussian curvature modulus by expanding the dynamic patch closure protocol to an equilibrium measurement based on externally confining fields, which after accounting for composition-curvature coupling can be generalized to the nontrivial case of lipid mixtures. In (2) Deserno explores the geometric nature of elastic forces that result from confining one-dimensional semi-flexible polymers to curved surfaces. Building on the case of a confining cylinder, which has both a continuous rotation and translation symmetry, the Euler-Lagrange equations for the shape are to be solved through a combination of analytical and numerical techniques, and the associated stresses and torques are identified. Moving on to a confining catenoid, translational symmetry is lost, as is generally a potential quadrature, but the new curvature gradients provide a host of new physics, linked to boundary conditions, curvature localization, and polymerization forces. Geodesics and asymptotic curves are studied as limiting cases for polymers with anisotropic elastic properties, and additional spontaneous curvature and twist render the filament an excellent continuum model for dynamin filaments. This allows the prediction of forces exerted by dynamin helices polymerizing around membrane necks, a process that is believed to underlie many cellular membrane fission events, but whose mechanistic underpinning is still not understood.


Grant
Agency: NSF | Branch: Cooperative Agreement | Program: | Phase: SPECIAL PROJECTS - CISE | Award Amount: 4.95M | Year: 2014

The eXpressive Internet Architecture (XIA) project aims to design, evaluate, and realize a future network architecture that provides inherent trustworthiness, supports long-term evolution of network use models and network technology, and addresses adoption through careful design of APIs and reasoned analysis of interactions (or tussle points) between multiple stakeholders. XIA offers intrinsic security by offering each entity the ability to unilaterally validate that it is communicating with the correct counterparties. XIA admits new usage models to evolve through the use of flexible addresses. This architectural flexibility also exposes new features, for example, enabling users to trade-off maintaining anonymity against maximizing efficiency in content retrieval. To date, the XIA team has developed an initial architecture; built prototype implementations of the XIA data plane, network protocol stack, and secure unicast communication using Scalability, Control and Isolation On next-generation Networks (SCION); engaged in a wide range of basic networking and security research; and investigated questions raised by future Internet architecture research.

In the next phase of research on XIA, principal investigators (PIs) advance the original vision of XIA by deepening the ongoing research thrusts, driven in part by research questions exposed during the initial investigation. The research agenda also includes an increased emphasis on control protocols based on a unified control plane architecture. The research will be driven through two network deployments that challenge the architectural framework: a vehicular network deployment in the city of Pittsburgh, and a video delivery environment spanning the U.S. These deployments also leverage and deepen the work on secure network operations, including providing a highly available infrastructure and secure authentication mechanisms. Finally, these deployments necessitate further research on building a robust XIA network, and establish best practices for using the XIA architecture, including support for mobility and a rich session layer. The research will continue to be informed by issues of deployability and economic viability, and to consider XIA as perceived by end users. The plan therefore includes research on governance, economic implications of high availability, and user studies that investigate the interplay between privacy, transparency and user control. Finally, the PIs will develop methods for performing a comparative evaluation of future Internet architectures. This evaluation methodology includes a definition of evaluation criteria, a discussion of the importance of direct and indirect evaluation, and evaluation using the XIA network environments and other use cases.

Intellectual Merit: The intellectual merit of the proposed work is an ambitious study of the effectiveness of a new architectural framework, evaluated in the context of two challenging network environments. The cross-cutting expertise of the team enables broad and deep investigation, starting from specification and implementation of many key network and security building blocks, up to application-specific methods delivering innovative solutions in the vehicular and video delivery environments. The evaluation spans multiple disciplines, ranging from low-level performance evaluation to reasoning about incentive-compatibility and economic benefits of the PIs approach to a quantification of user experience.

Broader Impacts: The project will establish a framework for evaluating network architectures that will be broadly applicable to FIA research. The PIs propose a variety of education and outreach activities to accompany the proposed research. Educational activities include new courses on clean-slate network design, and teaching and internship activities for under-represented and high school students. Broader outreach activity also includes a plan to involve and inform policymakers on the health of the current and future Internet led by a PI who formerly served at the FCC.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: SPECIAL PROJECTS - CISE | Award Amount: 2.87M | Year: 2013

Natural language privacy policies have become a de facto standard to address expectations of notice and choice on the Web. Yet, there is ample evidence that users generally do not read these policies and that those who occasionally do struggle to understand what they read. Initiatives aimed at addressing this problem through the development of machine implementable standards or other solutions that require website operators to adhere to more stringent requirements have run into obstacles, with many website operators showing reluctance to commit to anything more than what they currently do. This project offers the prospect of overcoming the limitations of current natural language privacy policies without imposing new requirements on website operators.

This frontier project builds on recent advances in natural language processing, privacy preference modeling, crowdsourcing, formal methods, and privacy interfaces to overcome this situation. It combines fundamental research with the development of scalable technologies to semi-automatically extract key privacy policy features from natural language website privacy policies and present these features to users in an easy-to-digest format that enables them to make more informed privacy decisions as they interact with different websites. Work in this project also involves the systematic collection and analysis of website privacy policies, looking for trends and deficiencies both in the wording and content of these policies across different sectors and using this analysis to inform ongoing public policy debates. An important part of this project is to work closely with stake holders in industry to enable the transfer of these technologies to industry for large-scale deployment.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: STATISTICS | Award Amount: 262.32K | Year: 2015

Data in high dimensional spaces are now very common. This project will develop methods for analyzing these high dimensional data. Such data may contain hidden structures. For example, clusters (which are small regions with a large number of points) can be stretched out like a string forming a structure called a filament. Scientists in a variety of fields need to locate these objects. It is challenging since the data are often very noisy. This project will develop rigorously justified and computationally efficient methods for extracting such structures. The methods will be applied to a diverse set of problems in astrophysics, seismology, biology, and neuroscience. The project will advance knowledge in several fields including computational geometry, astronomy, machine learning, and statistics.

Finding hidden structure is useful for scientific discovery and dimension reduction. Much of the current theory on nonlinear dimension reduction assumes that the hidden structure is a smooth manifold and is very restrictive. The data might be concentrated near a low dimensional but very complicated set, such as a union of intersecting manifolds. Existing algorithms, such as the Subspace Constrained Mean Shift exhibit erratic behavior near intersections. This project will develop improved algorithms for these cases. At the same time, contemporary theory breaks down in these cases and this project will develop new theory to address the aforementioned problem. A complete method (which will be called singular clusters) will be developed for decomposing point clouds of varying dimensions into subsets.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: COMM & INFORMATION FOUNDATIONS | Award Amount: 259.30K | Year: 2016

Cloud storage systems increasingly form the backbone of software services that underwrite everyday lives, serving a large array of businesses and forming an essential pillar of the economy. Given the sheer volume of interactions within these distributed systems, there arise a multitude of issues in building, maintaining and enhancing them. The most salient of these challenges are in the reliability, availability, consistency, confidentiality and privacy of data stored in these vast systems.

The proposed research is expected to have a significant impact on the manner in which cloud storage systems are designed and deployed. In these systems, storage node failures can have a significant impact on the efficiency of the overall system. This project enhances fault tolerant mechanisms to enable efficient recovery from failures, while augmenting the overall data availability and privacy offered by such systems. The research effort will advance the science of cloud computing by developing a new family of algorithms for distributed storage , and connect the advances to the significant industry needs in this topic. The research agenda will also be tightly integrated with education and outreach activities with direct involvement of underrepresented minorities, graduate, and undergraduate students.

Focusing on efficient maintenance of data with a range of desirable qualities, including mechanisms that ease data encoding, accessibility, updates, as well as privacy, the main objectives of this effort include (i) to develop coding schemes where a failed element can be regenerated with higher repair efficiencies from its local neighbors, (ii) to bring together the advantages of both local decodability and local repairability into one coding solution, (iii) to design mechanisms that provide low cost data updates in addition to efficient repair, (iv) to develop codes that can be resilient against failures with different scales/modalities, (v) to develop coding mechanisms taking advantage of implementation aspects of existing systems, and (vi) to develop coding schemes that enable users to access their data in a private manner. This effort addresses these challenges using a combination of tools from disciplines spanning coding theory, information theory, communications, as well as combinatorial and discrete mathematics.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: COMPUTING RES INFRASTRUCTURE | Award Amount: 542.37K | Year: 2013

The Speech Recognition Virtual Kitchen

Performing successful research on end-to-end speech processing problems requires the integration of many individual tools (e.g. for data cleaning, acoustic model training, language modeling, data analysis, real-time audio, decoding, parsing, synthesis, etc.). It is difficult for new researchers to get started in the field, simply because a typical lab environment consists of a hodgepodge of tools suited to a particular computing set-ups. This environment is hard to recreate, because few people are experts in the theory and practice of all these fields, and can debug and replicate experiments from scratch.

This research infrastructure project creates a kitchen environment based on Virtual Machines (VMs) to promote community sharing of research techniques, and provides solid reference systems as a tool for education, research, and evaluation. We liken VMs to a kitchen because they provide an environment into which one can install appliances (e.g., toolkits), recipes (scripts for creating state-of-the art systems using these tools), and ingredients (spoken language data). The kitchen even holds reference dishes in the form of complete experiments with baseline runs, log-files, etc., together with all that is needed to recreate and modify them.

The project is developing a community and repository by (a) building pilot VMs, (b) engaging the community in using and continuing to develop them on its own, and (c) evaluating the impact of providing VMs for education and research. We envision researchers as well as students downloading a VM, reproducing the baseline experiment, implementing changes, posting their results in the community, discussing with other users who have worked on the same VM, merging improvements back into the VM, which get re-distributed, and finally publishing easily reproducible results. Work with curriculum and project development will support the creation of engaging activities to specifically encourage students at undergraduate and graduate levels.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: PETASCALE - TRACK 1 | Award Amount: 40.00K | Year: 2016

The Hubble Deep/Ultra Deep fields are iconic, and almost certainly the
most studied space telescope observations. Upcoming observations from an enormous range of
current and planned instruments, from the successor to Hubble, The James Webb Telescope, to
the Large Synoptic Survey Telescope (LSST), to the WFIRST Satellite will also have the
opportunity to reach galaxies and black holes at the same extreme magnitudes or deeper, but
over unprecedentedly wide fields. This will open up the study of the highest redshift galaxies
and quasars to the type of statistical studies that have made modern cosmology a precision
science. This project aims to compute theoretical predictions to make contact with current
and upcoming observations of the high-redshift universe using the Blue Waters supercomputer.

The project will extend the BlueTides simulation, with an unprecedented
volume and resolution, to cover the evolution of the first billion years of cosmic history.
The goal is to significantly increase the scientific impact of this calculation to the community.
Importantly, the project will attempt to make contact
with observations of quasars, which have not been discovered at redshift greater than 7 (while
the simulation now has run to redshift equal 8). In addition the project will be using BlueTides as
a path-finder for developing methods/calculations for future cosmological hydrodynamical simulations
of galaxy formation with volumes and resolutions suitable for creating mocks for next generation
surveys. The impact of the proposed work will extend way beyond BlueTides. Large hydrodynamical simulations
will be more and more useful in all stages of major observational projects in astrophysics
and cosmology. For example, a simulation that covers a significant fraction of the entire
observable universe with BlueTides resolution and runs to the present day (an epoch which
is fully dominated by the hydrodynamic computations) will be needed for LSST.

The project will establish a theoretical framework for understanding the role of galaxies
in the evolution of the universe at high redshifts. Different communities
of scientists are interested in the behavior history of quasars and galaxy assembly, including
cosmologists, the galaxy evolution community and high energy astrophysicists, so the results
would have a wide impact across many different scientific communities.

Additionally, the image and catalog generation, and
database techniques developed by the project will
strengthen the project already on-going synergistic activities with computer science, machine learning
and statistics. Furthermore, the project will have a strong education component by involving undergraduate
and graduate students in this research. Finally, the project propose to perform outreach using
the visualization and interactive Gigapan software.


Grant
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 15.00K | Year: 2015

Scientific findings and innovations have shaped American life in countless ways, but not all scientific advances have received a warm reception by the public. This project investigates the factors that predict public acceptance or rejection of scientific evidence, focusing on ?sophisticated rejectionists,? individuals who possess the ability to evaluate scientific evidence critically, but still reject the scientific consensus on key issues. The results of the proposed research will identify situations where sophisticated rejection of science is likely, understand how people maintain beliefs contrary to the scientific consensus (e.g., childhood vaccinations) and suggest ways to communicate about science so that it receives a fair hearing.

Scientists take rigorous precautions to ensure the validity of their research and the accuracy of its reporting. However, the public may reject the scientific consensus on an issue (e.g., genetically modified foods) based on factors other than those employed by scientists. This research program examines one possible contributor to such rejection: individuals may lack the skills needed to understand and evaluate scientific evidence. In earlier work, the researcher developed and validated a novel scale measuring nonscientists? scientific reasoning skills (SRS), defined as those needed to assess the validity of scientific results. The present research uses this scale to understand a group of individuals identified in the research to date: ?sophisticated rejectionists,? defined as individuals who have high SRS scores yet still reject the scientific consensus on key issues. The first study will analyze data from the National Center for Science and Engineering Statistics? Survey of Public Attitudes Toward and Understanding of Science and Technology, looking for evidence of such sophisticated rejectionists. The second study will investigate how they reject science. Do they fail to apply their skills to dissect spurious scientific arguments or is their interpretation of scientific evidence biased by their belief in claims that most scientists would consider false or incomplete? Two additional studies will test strategies to reduce sophisticated rejection of science by varying how science is communicated. As a whole, this research furthers scientific understanding of how nonscientists evaluate, and sometimes reject, scientific evidence.


Grant
Agency: NSF | Branch: Continuing grant | Program: | Phase: | Award Amount: 589.41K | Year: 2015

CIF: Medium: Data Science: Analytics for Unstructured and Distributed Data
José M F Moura and Soummya Kar

Abstract

There has been a perfect storm of convergent technologies leading to an onslaught of data in digital format, ready to be acquired, stored, accessed, transmitted, and processed. This research develops quantitative analytics that are appropriate for data arising in many novel areas like social networks or urban environments, outside traditional engineering or science. This data is unstructured?it is no longer a single time series or a single image and cannot be naturally arranged in a vector or table. The data is distributed?it originates from many different agents, possibly scattered over a large physical space (e.g., a metro area).

To process unstructured and distributed Big Data, the research extends traditional signal processing methods to distributed signal processing on gra