Time filter

Source Type

Nagoya-shi, Japan

The Toyota Technological Institute is a university located in Nagoya, Japan. Founded in 1981 by a large endowment from Toyota Motors Corporation, it originally only accepted students with some industrial work experience.TTI has a School of Engineering, a Master's Program and a Doctoral Program. The programs consist of three areas of coursework: Mechanical Systems Engineering, Electronics & Information Science, and Materials Science & Engineering. In 2003 Toyota also opened the Toyota Technological Institute at Chicago, jointly with the University of Chicago. This campus is mainly for the Ph.D students, studying Machine Learning, Algorithms & Complexity, Computer Vision, Speech Technologies and Computational Biology.Although TTI is a new and tiny university, it has been rapidly growing its reputation, as TSU ranked TTI as the 5th best Japanese university in 2010 and 4th in 2011. In this ranking, TTI has a best employment rate among all Japanese Universities.In 2012, TTI was ranked 1st in Asia in terms of average number of publication per faculty by the QS World University Rankings. Wikipedia.

Awano H.,Toyota Technological Institute
Journal of Magnetism and Magnetic Materials | Year: 2015

Current driven magnetic domain wall (DW) motions of ferri-magnetic TbFeCo wires have been investigated. In the case of a Si substrate, the critical current density (Jc) of DW motion was successfully reduced to 3×106 A/cm2. Moreover, by using a polycarbonate (PC) substrate with a molding groove of 600 nm width, the Jc was decreased to 6×105 A/cm2. In order to fabricate a logic in memory, a current driven spin logics (AND, OR, NOT) have been proposed and successfully demonstrated under the condition of low Jc. These results indicate that TbFeCo nanowire is an excellent candidate for next generation power saving memory and logic. © 2015 The Author. Source

Matsui H.,Toyota Technological Institute
IEEE Transactions on Information Theory | Year: 2014

In this paper, we establish a lemma in algebraic coding theory that frequently appears in the encoding and decoding of, e.g., Reed-Solomon codes, algebraic geometry codes, and affine variety codes. Our lemma corresponds to the nonsystematic encoding of affine variety codes, and can be stated by giving a canonical linear map as the composition of an extension through linear feedback shift registers from a Gröbner basis and a generalized inverse discrete Fourier transform. We clarify that our lemma yields the error-value estimation in the fast erasure-And-error decoding of a class of dual affine variety codes. Moreover, we show that systematic encoding corresponds to a special case of erasure-only decoding. The lemma enables us to reduce the computational complexity of error-evaluation from O(n^{3}) using Gaussian elimination to O(qn^{2}) with some mild conditions on n and q , where n is the code length and q is the finite-field size. © 1963-2012 IEEE. Source

Tiwari J.N.,Pohang University of Science and Technology | Tiwari R.N.,Toyota Technological Institute | Kim K.S.,Pohang University of Science and Technology
Progress in Materials Science | Year: 2012

One of the biggest challenges of 21st century is to develop powerful electrochemical energy devices (EEDs). The EEDs such as fuel cells, supercapacitors, and Li-ion batteries are among the most promising candidates in terms of power-densities and energy-densities. The nanostructured materials (NSMs) have drawn intense attention to develop highly efficient EEDs because of their high surface area, novel size effects, significantly enhanced kinetics, and so on. In this review article, we briefly introduce general synthesis, fabrication and their classification as zero-dimensional (0D), one dimensional (1D), two-dimensional (2D) and three-dimensional (3D) NSMs. Subsequently, we focus an attention on recent progress in advanced NSMs as building blocks for EEDs (such as fuel cells, supercapacitors, and Li-ion batteries) based on investigations at the 0D, 1D, 2D and 3D NSMs. © 2011 Elsevier Ltd. All rights reserved. Source

Agency: NSF | Branch: Standard Grant | Program: | Phase: ADVANCES IN BIO INFORMATICS | Award Amount: 557.03K | Year: 2016

Proteins play fundamental roles in all biological processes. Complete description of protein structures and functions is a fundamental step towards understanding biological life and has various applications. Millions of protein sequences are available, but a majority of them have no experimentally-solved structures and functions. This project aims to greatly improve RaptorX, a fully-automated web server for computational prediction of protein structure and function, with the goal to deliver a long-term sustainable web portal to facilitate transformative research in biology. This web portal shall benefit a broad range of biological and biomedical applications, such as genome annotation, understanding of disease processes, drug design, precision medicine and even biomaterial and bio-energy development. The results will be disseminated to the broader community through a variety of venues: web servers, standalone software, publications and talks. Since late in 2011, RaptorX has served >25,000 worldwide users including middle- and high-school students. The standalone programs have been downloaded by >1500 worldwide users. After this project is fulfilled, RaptorX will contribute much more to the broader community. Students involved in this project will receive training in the intersection of computer science, molecular biology, biophysics, and biochemistry. Undergraduate and underrepresented students will be recruited through summer intern programs and collaborators. The research results will be integrated into course materials and used in the Illinois online bioinformatics program.

The RaptorX web server was originally developed for only template-based protein modeling. This project will transform RaptorX by first developing a few novel and powerful deep learning (e.g., Deep Conditional Convolutional Neural Fields) and structure learning (e.g., group graphical lasso) methods to significantly improve the accuracy of protein structure and functional prediction and then conducting an efficient implementation. The resultant RaptorX will be able to perform much more accurate prediction of protein secondary and tertiary structure, solvent accessibility and disordered regions, and the quality of a theoretical protein 3D model (in the absence of natives). This project will also expand the RaptorX server to perform contact prediction and contact-assisted protein folding for proteins without good templates. The RaptorX web server is available at http://raptorx.uchicago.edu, from which users can also download the standalone programs.

Agency: NSF | Branch: Standard Grant | Program: | Phase: ROBUST INTELLIGENCE | Award Amount: 194.61K | Year: 2016

Vision is a valuable sensing modality because it is versatile. It lets humans navigate through unfamiliar environments, discover assets, grasp and manipulate tools, react to projectiles, track targets through clutter, interpret body language, and recognize familiar objects and people. This versatility stems from low-level visual processes that somehow produce, from ambiguous retinal measurements, useful intermediate representations of depth, surface orientation, motion, and other intrinsic scene properties. This project establishes a mathematical and computational foundation for similar low-level processing in machines. The key challenge it addresses is how to usefully encode and exploit the fact that, visually, the world exhibits substantial intrinsic structure. By advancing understanding of low-level vision in machines, this project makes progress toward computer vision systems that can compare to vision in humans, in terms of accuracy, reliability, speed, and power-efficiency.

This research revisits low-level vision, and develops a comprehensive framework that possesses a common abstraction for information from different optical cues; the ability to encode scene structure across large regions and at multiple scales; implementation as parallel and distributed processing; and large-scale end-to-end learnability. The project approaches low-level vision as a structured prediction task, with ambiguous local predictions from many overlapping receptive fields being combined to produce a consistent global scene map that spans the visual field. The structured prediction models are different from those used for categorical tasks such as semantic segmentation, because they are specifically designed to accommodate the distinctive requirements and properties of low-level vision: continuous-valued output spaces; ambiguities that may form equiprobable manifolds; extreme scale variations; and global scene maps with higher-order piecewise smoothness. By strengthening the computational foundations of low-level vision, this project strives to enable many kinds of vision systems that are more efficient and more versatile, and it strives to have impacts across the breadth of computer vision.

Discover hidden collaborations