CDS is engaged in cutting-edge research programs in areas relating to computational and data science, computing and data systems, and their applications. The research focus of CDS is broadly classified into Computer and Data Systems stream and Computational Science stream, with the labs within each stream described below.
Please visit the research areas page to understand common themes that cut across these labs.
Computer and Data Systems Labs
|Lab Name||Faculty (Convener)||Description|
|Computer Aided Design (CAD) Lab||S. K. Nandy||The lab is engaged in research related to Runtime Reconfigurable System-on-Chip (SoC) architectures, modeling of massively parallel reconfigurable silicon cores, compilation techniques for reconfigurable silicon cores, large scale simulations on massively parallel and distributed SoC architectures, and application synthesis on Runtime Reconfigurable SoCs.|
|Database Systems Lab||Jayant R. Haritsa||The lab works on the design, evaluation, testing and tuning of the internals of industrial-strength database engines, covering topics such as query optimization and processing, transaction processing mechanisms, data security and privacy, and operator design for modern architectures.|
|Middleware and Runtime Systems (MARS) Lab||Sathish Vadhiyar||This lab focuses on research in High Performance Computing (HPC) involving challenging parallel applications (large-scale, long-running, dynamic, irregular, multi-component etc.) and challenging parallel systems, namely, GPUs, state-of-art supercomputers and grids.
Application Frameworks: This work involves building generic frameworks, runtime strategies, user interfaces and abstractions, and programming models for applications on accelerator-based systems and Torus networks. We focus on three classes of applications, namely, irregular, multi-physics and climate/weather modeling applications.
Middleware for Batch Systems: Our research aims to build a middleware framework that interfaces between the users and the batch queues and systems. The middleware will have prediction techniques that predict queue waiting times and the execution times incurred by the parallel jobs submitted to the batch queues, and scheduling strategies that use these prediction techniques to assign the appropriate batch queue and number of processors for job execution with the aim of reducing the turnaround times of the users and increasing the throughput of the system.
Fault Tolerance: Use of replication for fault tolerance in HPC applications.
|Video Analytics Lab (VAL)||R. Venkatesh Babu||Home land security has become most important international effort to prevent various physical security threat nowadays. Video analytics is the brain behind the success of such a large scale system. The objective of the lab is to develop robust and reliable algorithms for semantic understanding of the visual content. Current research at Video Analytics Lab focuses on object detection, human activity analysis, visual tracking, compressed domain video analytics, visual information retrieval medical image analysis etc.|
|Distributed Research on Emerging applications and Machines Lab (DREAM:Lab)||Yogesh Simmhan||The Distributed Research on Emerging Applications and Machines Lab (DREAM:Lab) focuses on distributed systems research that enables the effective and efficient use of emerging distributed data and computing systems, using scalable software architectures, innovative programming and data abstractions, and algorithms for optimal distributed execution. We explore Big Data platforms, Cloud and Edge computing, Distributed Graph Processing, Distributed video analytics, and Internet of Things (IoT) applications for UAV/UGVs. These support data intensive scientific and engineering applications, which can lead to transformative advances to society.
|Machine And Language Learning (MALL) Lab||Partha Pratim Talukdar||Currently, the lab is focused on extraction of knowledge from Web-scale datasets (primarily textual) and application of such knowledge in various real-world applications. One of the primary goals of the lab is to overcome the knowledge-bottleneck problem in intelligent decision making. Research carried out in the lab spans the areas of Machine Learning and Natural Language Processing.|
|Visual Computing Lab (VCL)||Anirban Chakraborty||Members of the Visual Computing Lab (VCL) seeks solutions to computer vision and machine learning problems arising in numerous application areas involving visual data analytics. These include data association problems over large graphs such as person re-identification in camera networks, video surveillance problems such as face recognition, human activity analysis, abnormality detection etc. and biological and medical image analysis.|
Computational Science Labs
|Scientific Computation Lab||Soumendyu Raha||Current research interest is in computational methods for simulation, optimization and control of constrained and stochastic dynamical systems, and application to problems in Mechanics and Biochemical Kinetics. There is also an interest in codesign of numerical algorithm and architectural elements of accelerators toward efficient and high performance implementation of the computational methods for dynamical systems.|
|Biomolecular Computation Lab||Debnath Pal||Computational Biology and Bioinformatics is an interdisciplinary subject area that seeks to answer important questions in biological science exploiting multi-disciplinary expertise from various areas like Physics, Chemistry, Mathematics, Biology, Computer science and engineering. Biological questions are the central point in these studies and interdisciplinary expertise is leveraged to address specific questions. The scope of the questions could range from simple, such as, what is a protein sequence coded by a gene to questions on complex cellular mechanism and signal transductions. Given the diversity and wide range of areas that encompass this subject, each researcher focuses on a few or related questions of his or her interest. For example Prof. Debnath Pal’s laboratory currently focuses on the theme of Sequence→Structure→Function relations in biomolecules, protein in particular. The final goal is to understand the function of biomolecules in molecular and cellular context. Given that function of any biomolecule is governed by its physico-chemical features and environmental contexts, his lab approaches the problem from various angles. For example, addressing the question from the perspective of first principle, Prof. Pal’s lab is studying protein function through its correlation with molecular dynamics. The simple idea that is being investigated is that proteins are flexible molecules ─ can we find specific functional modes of motion to understand what and how proteins perform their function? The same problem is also being approached from the experimental side on a macroscopic scale through Systems Biology approaches using genomics, transcriptomics, proteomics, and metabolomics. Here molecules are being investigated to characterize them for identification, followed by understanding their modification and/or concentration in cells. The data is being used to develop models and hypothesis for biological function. On computer science side, new algorithms are being developed for better and more efficient answer to the said problems.|
|Computational and Statistical Physics Lab (CSPL)||Murugesan Venkatapathi||A part of this group studies behavior of light and light-matter interactions using analytical and numerical models. Another part works on developing computational formulations for physics and matrix algebra.|
|Structural Biology and Bio-computing Lab||K. Sekar||The research group focuses on solving three-dimensional crystal structures
of protein molecules using X-ray crystallography and molecular dynamics
simulations. Further, we are also interested in data mining of protein
sequences and structures.
|Medical Imaging Group||Phaneendra K. Yalavarthy||The Medical Imaging Group (MIG) focuses on developing novel computational methods in medical imaging. The group works on Medical Image Reconstruction, High Performance Computing in Medical Image Processing, Biomedical Optical Image Reconstruction (Diffuse Optical Tomography and Photoacoustic Tomography), and Neuroimaging (fMRI analysis and atlas/template creation). The group has active collaborations with NIMHANS, Bangalore.|
|Computational Mathematics||Sashikumaar Ganesan||The research group focus on the development and advancement of robust numerical (finite element) methods for solving partial differential equations (PDEs) which describe incompressible fluid flows and species concentration and/or energy in complex systems. The group has ongoing collaboration with several academic in Germany. The group works on – but is not limited to – implementing efficient algorithms for high-performance computing. ParMooN is an in-house open source finite element package of this group.|
|Quantifying Uncertainty in Engineering, Science and Technology (QUEST) Lab||Deepak Subramani||The goal of our research is to build holistic science-based data-driven computational solutions to complex engineering and environmental problems. Example applications include climate change, cyclone predictions, coastal hazard management, and optimal vehicle routing. Pursuant to our goal, we develop and apply fundamental theories, numerical schemes and software systems. We invite students who are interested in any of the following topics: numerical solution of stochastic partial differential equations, uncertainty quantification, Bayesian and deep learning of dynamical systems, fluid dynamics of the atmosphere and oceans, and path planning of autonomous vehicles in dynamic environments.|
|FLow Analysis and Multi-Physics-simulations at Extreme-scale (FLAME) Lab||Konduri Aditya||The research group focuses on the simulation and analysis of multi-scale multi-physics fluid flow problems that leverage high performance computing (HPC) platforms. Specifically, the work would involve development of HPC centric numerical methods and algorithms for solving partial differential equations, application of machine learning methods to analyze and model data generated from simulations, and investigation of turbulent flow problems that arise in combustion systems, high-speed aerodynamics and environmental flows.|
|Materials-physics & Algorithmic Techniques Research In eXtreme-computing (MATRIX) Lab||Phani Motamarri||MATRIX lab's research is centered around developing techniques which are aimed at pushing the boundaries of the current predictive capabilities of computation based design of materials, with an objective of addressing complex material science problems than possible heretofore. In particular, the research involves -- (i) development and efficient implementation of mathematical techniques and HPC driven computational algorithms that can leverage the latest heterogeneous parallel computing architectures and future exa-scale machines (CPU+GPU) for quantum mechanical modeling of materials, (ii) employing machine learning strategies towards accelerating these computational methods for material modeling (iii) harnessing these novel computational capabilities to address challenging material modeling problems which can provide deeper insights into various aspects of material properties at the nano-scale, thereby informing higher-scale models for accurate prediction of macroscopic material properties. Research at MATRIX lab is highly interdisciplinary and combines deep ideas from condensed matter physics, materials science, mechanics of solids, adaptive finite-element methods, numerical methods, machine learning and heavy dose of high performance computing (HPC) .|
|Algorithmic Techniques for Computational Genomics (ATCG) Lab||Chirag Jain||The ATCG lab develops cutting-edge computational and mathematical techniques for data-intensive problems in molecular biology. These techniques find applications in areas such as human genetics, infectious diseases and cancer research. Interpreting DNA molecules and genomes using tera or peta bytes of raw sequencing data has been the key to understanding biological role of genetic mutations. Our research within this area builds on and extends nearest-neighbour search, string and graph algorithms, along with their parallelisation on modern high performance computing platforms. Often, our projects also involve close collaboration with domain experts in biology.