Dwarka, India

Indraprastha Institute of Information Technology, Delhi is an autonomous university in Delhi, India. Wikipedia.

Time filter

Source Type

Agency: Cordis | Branch: FP7 | Program: NoE | Phase: SEC-2012.7.4-2 | Award Amount: 5.55M | Year: 2014

VOX-Pol is a 60-month, 5\ million project that integrates the worlds leading researchers and research groups in Violent Online Political Extremism (VOPE), to include those researching the intersection of terrorism and the Internet (incl. violent jihadists, violent separatists, etc.), the online activities of the extreme Right, the potential for violent online radicalisation, etc., in order to: 1. Create a sustainable critical mass of innovative activity among what is currently a burgeoning, but fragmented group of researchers and research topics. 2. Ensure that EU and MS strategies and policies targeting VOPE are based on concrete evidence, experience, and knowledge about the contours and workings of VOPE and thus increasing their likelihood of success. VOX-Pol will: Integrate and network the research activities of those, within the EU and globally, working in the area of VOPE Create and develop long-term relationships between established national research groups, new researchers and research groups, security practitioners, the Internet industry, civil society, and policymakers leading to the development of a multi-disciplinary Virtual Centre of Excellence for Research in VOPE Be based on collaborative research among partners both within and outside the Network, with all contributing towards the development of an archive of politically extreme Internet-based content and a related URL database, which will be the basis for joint research activity, the development of new analytical tools and methodologies, teaching and training, and dissemination activities Raise awareness of the challenges of research and decision-making in this area by exploring the interplay of e-research ethics, privacy, surveillance, freedom of speech, and practices of and responses to VOPE Influence research agendas on the European and world stages in key aspects of VOPE Inform policy agendas on national, European, and international levels in key aspects of responses to VOPE.

Majumdar A.,Indraprastha Institute of Information Technology | Ward R.K.,University of British Columbia
Biomedical Signal Processing and Control | Year: 2014

This work addresses the problem of reconstructing EEG signals from lower dimensional projections. Unlike previous studies, we propose to reconstruct the EEG signal using an analysis prior formulation. Moreover we use the inter-channel correlation while reconstruction which leads to a row-sparse analysis prior multiple measurement vector (MMV) recovery problem. To improve the reconstruction, we formulate the recovery as a non-convex optimization problem. Such a non-convex row-sparse MMV recovery problem had not been encountered before; this work derives an efficient algorithm to solve it. The proposed reconstruction technique is compared with state-of-the-art methods and we find that our technique yields significant improvement over others. © 2014 Elsevier Ltd. All rights reserved.

Sureka A.,Indraprastha Institute of Information Technology | Jalote P.,Indraprastha Institute of Information Technology
Proceedings - Asia-Pacific Software Engineering Conference, APSEC | Year: 2010

We present an approach to identify duplicate bug reports expressed in free-form text. Duplicate reports needs to be identified to avoid a situation where duplicate reports get assigned to multiple developers. Also, duplicate reports can contain complementary information which can be useful for bug fixing. Automatic identification of duplicate reports (from thousands of existing reports in a bug repository) can increase the productivity of a Triager by reducing the amount of time a Triager spends in searching for duplicate bug reports of any incoming report. The proposed method uses character N-gram-based model for the task of duplicate bug report detection. Previous approaches are word-based whereas this study investigates the usefulness of low-level features based on characters which have certain inherent advantages (such as natural-language independence, robustness towards noisy data and effective handling of domain specific term variations) over word-based features for the problem of duplicate bug report detection. The proposed solution is evaluated on a publicly-available dataset consisting of more than 200 thousand bug reports from the open-source Eclipse project. The dataset consists of ground-truth (pre-annotated dataset having bug reports tagged as duplicate by the Triager). Empirical results and evaluation metrics quantifying retrieval performance indicate that the approach is effective. © 2010 IEEE.

Parkash A.,Indraprastha Institute of Information Technology | Parikh D.,Toyota Technological Institute
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2012

Traditional active learning allows a (machine) learner to query the (human) teacher for labels on examples it finds confusing. The teacher then provides a label for only that instance. This is quite restrictive. In this paper, we propose a learning paradigm in which the learner communicates its belief (i.e. predicted label) about the actively chosen example to the teacher. The teacher then confirms or rejects the predicted label. More importantly, if rejected, the teacher communicates an explanation for why the learner's belief was wrong. This explanation allows the learner to propagate the feedback provided by the teacher to many unlabeled images. This allows a classifier to better learn from its mistakes, leading to accelerated discriminative learning of visual concepts even with few labeled images. In order for such communication to be feasible, it is crucial to have a language that both the human supervisor and the machine learner understand. Attributes provide precisely this channel. They are human-interpretable mid-level visual concepts shareable across categories e.g. "furry", "spacious", etc. We advocate the use of attributes for a supervisor to provide feedback to a classifier and directly communicate his knowledge of the world. We employ a straightforward approach to incorporate this feedback in the classifier, and demonstrate its power on a variety of visual recognition scenarios such as image classification and annotation. This application of attributes for providing classifiers feedback is very powerful, and has not been explored in the community. It introduces a new mode of supervision, and opens up several avenues for future research. © 2012 Springer-Verlag.

Majumdar A.,Indraprastha Institute of Information Technology
Magnetic Resonance Imaging | Year: 2015

In blind compressed sensing (BCS), both the sparsifying dictionary and the sparse coefficients are estimated simultaneously during signal recovery. A recent study adopted the BCS framework for recovering dynamic MRI sequences from under-sampled K-space measurements; the results were promising. Previous works in dynamic MRI reconstruction showed that, recovery accuracy can be improved by incorporating low-rank penalties into the standard compressed sensing (CS) optimization framework. Our work is motivated by these studies, and we improve upon the basic BCS framework by incorporating low-rank penalties into the optimization problem. The resulting optimization problem has not been solved before; hence we derive a Split Bregman type technique to solve the same. Experiments were carried out on real dynamic contrast enhanced MRI sequences. Results show that, with our proposed improvement, the reconstruction accuracy is better than BCS and other state-of-the-art dynamic MRI recovery algorithms. © 2014 Elsevier Inc.

Sureka A.,Indraprastha Institute of Information Technology
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2012

Bug reports in widely used defect tracking systems contains standard and mandatory fields like product name, component name, version number and operating system. Such fields provide important information required by developers during bug fixing. Previous research shows that bug reporters often assign incorrect values for such fields which cause problems and delays in bug fixing. We conduct an empirical study on the issue of incorrect component assignments or component reassignments in bug reports. We perform a case study on open-source Eclipse and Mozilla projects and report results on various aspects such as the percentage of reassignments, distribution across number of assignments until closure of a bug and time difference between creation and reassignment event. We perform a series of experiments using a machine learning framework for two prediction tasks: categorizing a given bug report into a pre-defined list of components and predicting whether a given bug report will be reassigned. Experimental results demonstrate correlation between terms present in bug reports (textual documents) and components which can be used as linguistic indicators for the task of component prediction. We study component reassignment graphs and reassignment probabilities and investigate their usefulness for the task of component reassignment prediction. © 2012 Springer-Verlag.

Mohsina M.,Indraprastha Institute of Information Technology | Majumdar A.,Indraprastha Institute of Information Technology
Biomedical Signal Processing and Control | Year: 2013

This paper deals with the problem of tele-monitoring EEG signals. In EEG tele-monitoring system, the integral step is to compress the signals in computationally efficient manner so that they can be transmitted over a limited bandwidth. In such a situation a Compressed Sensing (CS) framework for compressing and recovering the signals is the most viable approach. Previously the well known synthesis prior formulation is used for reconstruction. For the first time in this work, we show that the lesser known analysis prior formulation is a more appropriate way to frame the reconstruction problem. We show that our method yields better results than the previous synthesis prior formulation. © 2013 Elsevier Ltd.

Ram S.S.,Indraprastha Institute of Information Technology
IEEE National Radar Conference - Proceedings | Year: 2015

Current through-wall radar implementations yield top-view images of human activities. However, radar operators may find frontal images of humans easier to interpret since they resemble human perspectives. High-resolution frontal images are challenging to generate since they require large radar apertures and high carrier frequencies. We propose a methodology to generate high-resolution frontal images of moving humans using a Doppler radar with a limited size aperture operating at a low carrier frequency. The target is imaged using three dimensional Fourier processing along azimuth, elevation and Doppler. The additional Doppler dimension enables us to reduce the resolution requirements along the cross-range dimensions in terms of aperture size and number of elements. Additionally, we analyze the effect of Doppler resolution, the orientation and range of the human subject with respect to the radar for three different human activities - running, skipping and crawling. © 2015 IEEE.

Majumdar A.,Indraprastha Institute of Information Technology
Magnetic Resonance Imaging | Year: 2013

In this paper we address the problem of dynamic MRI reconstruction from partially sampled K-space data. Our work is motivated by previous studies in this area that proposed exploiting the spatiotemporal correlation of the dynamic MRI sequence by posing the reconstruction problem as a least squares minimization regularized by sparsity and low-rank penalties. Ideally the sparsity and low-rank penalties should be represented by the l0-norm and the rank of a matrix; however both are NP hard penalties. The previous studies used the convex l1-norm as a surrogate for the l0-norm and the non-convex Schatten-q norm (0

Rastogi A.,Indraprastha Institute of Information Technology
Proceedings - International Conference on Software Engineering | Year: 2015

Software project performance largely depends on the software development team. Studies have shown that interest and activity levels of contributors at any time significantly affect project success measures. This dissertation provides suggestions to enhance contributors' performance and participation intentions to help improve project performance. To do so, we mine historical data in software repositories from a two-pronged approach: 1) To assess contributors' performance to identify strengths and areas of improvement. 2) To measure the influence of factors on contributors' participation and performance, and provide suggestions that help advance contributor's engagement. The methodology used in this study leverage empirical techniques, both quantitative and qualitative, to conduct the analysis. We believe that the insights presented here will help contributors improve their performance. Also, we expect managers and business analysts to benefit from the suggestions to revise factors that negatively influence contributors' engagement and hence improve project performance. © 2015 IEEE.

Loading Indraprastha Institute of Information Technology collaborators
Loading Indraprastha Institute of Information Technology collaborators