Industry Research and User Symposium (IRUS) brings together solution providers and users of HPC, cloud and data platforms, providing a venue where they can share their challenges and successes and discuss relevant technology issues related to computational and data sciences. Previous IRUS sessions (2017, 2016 links) had tremendous participation from conference attendees and included senior speakers from the industry like Amazon, Flipkart, Intel, Microsoft, Shell and Xilinx and academia like CERN, IISC and TIFR.

Session 1: Convergence of HPC with AI

Date and Time: December 17th, 2018: 2:00 PM – 4:00 PM.

Speaker Bio: Rajkumar Kettimuthu received the B.E. degree from Anna University, Chennai, India, and an M.S. and Ph. D. from the Ohio State University, all in Computer Science and Engineering. He is a Computer Scientist in the Data Science and Learning Division at Argonne National Laboratory. He is also a Senior Scientist in the Consortium for Advanced Science and Engineering at The University of Chicago. His research is focused on advancing distributed data intensive science through innovations and intelligence in networks, system software and applications. Wide area data transfer solution developed by him and his colleagues at Argonne has become the de facto method for file transfers in many science environments. With more than 60000 installations in six continents, their tool performs more than 50 million file transfers and move more than 5 Petabytes of data every day. Currently, he and his colleagues are working on incorporating learning and intelligence into both the science HPC infrastructure and the science workflows. He has co-authored more than 100 peer-reviewed articles many of which appeared in premier journals and conferences in computer science. He is a recipient of the prestigious R&D 100 award. He is a distinguished member of ACM and a senior member of IEEE.

Talk Abstract: Rajkumar Kettimuthu received the B.E. degree from Anna University, Chennai, India, and an M.S. and Ph. D. from the Ohio State University, all in Computer Science and Engineering. He is a Computer Scientist in the Data Science and Learning Division at Argonne National Laboratory. He is also a Senior Scientist in the Consortium for Advanced Science and Engineering at The University of Chicago. His research is focused on advancing distributed data intensive science through innovations and intelligence in networks, system software and applications. Wide area data transfer solution developed by him and his colleagues at Argonne has become the de facto method for file transfers in many science environments. With more than 60000 installations in six continents, their tool performs more than 50 million file transfers and move more than 5 Petabytes of data every day. Currently, he and his colleagues are working on incorporating learning and intelligence into both the science HPC infrastructure and the science workflows. He has co-authored more than 100 peer-reviewed articles many of which appeared in premier journals and conferences in computer science. He is a recipient of the prestigious R&D 100 award. He is a distinguished member of ACM and a senior member of IEEE.

Speaker Bio: Jeff is a HPC specialist with over 25 years of experience in developing, tuning and  porting scientific codes and architecting HPC solutions. Jeff’s primary area of expertise is in CFD and NWP, having previously worked at the New Zealand Oceanographic Institute (now NIWA), Toyota Motor Corporation, and on FEA/CFD analysis for America’s cup class yachts for Team New Zealand. Prior to joining Nvidia, Jeff worked for SGI for 16 years in Asia, Before that, he worked for various Post Production companies in his native New Zealand as a Visual Effects artist, technical director, and software development roles. Jeff holds a post graduate diploma from the University of Auckland in Computer Science, specializing in Parallel programming and Computer graphics.

Talk Abstract: Learn about how AI is helping to find anomaly detection in climate change. You will see case study on how neural networks has helped getting better detection accuracy over traditional methods.

Speaker Bio: PhD student at CSIR-Institute of Genomics and Integrative Biology, New Delhi. I closely work with clinicians at AIIMS, New Delhi, for my PhD work on “critical event prediction in in-patient and out-patient pediatric healthcare”. I have published 5 peer reviewed papers in the field of computational medicine and healthcare informatics. I also have developed an android app “AnthroCal” for child growth assessment for AIIMS New Delhi. I am a recipient of CSIR-GATE fellowship for my PhD work. Previously I received Xerox International Travel Grant for representing my work on prediction of asthma exacerbation in children at European Association of systems medicine conference. Apart from the work, I enjoy playing Table Tennis, chess & science philosophy discussions over the tea table.

Talk Abstract: Critical events are the stages which lead to potentially irreversible outcomes such as organ dysfunction and mortality in the patients admitted to Intensive care unit of a hospital. Every hour of delay in treatment is associated with the increased risk of mortality. Shock and Sepsis are the two such most common complications in ICU, accounting for almost 30% incidence and 20 to 30% of all hospital mortality. Early prediction can enable timely treatment decision by clinical management thus, prevents the future risk of mortality and improves the patient care. The prediction of these events involves identification of early pattern in the patient data which can predict before a window of time in which interventions can play their role. For the prediction of shock, we developed an automated pipeline which captures the body temperature gradients to predict the hemodynamic shock and circulatory failures. Secondly for the prediction of sepsis, we utilised the patient physiological markers, captured as time-series. We generated a 3 hour early prediction sepsis cohort of multivariate time-series from 2168 patients; identified using sequential organ dysfunction assessment scores (SOFA scores).   We then learnt the time-series image representation using Hilbert space filling curve which is a form of a fractal space filling curves on this cohort. Finally we used the time-series image representation for early prediction using convolutional neural networks.  We used various state-of-art deep-learning architectures such as Very Deep Convolutional Neural Network and Inception-ResNet Models for the prediction of sepsis at 3 hour prior to its onset.

Speaker Bio: Yang Juntao is currently working as a Solution Architect at Nvidia. He did his MSc from National University of Singapore. His focus areas include looking at Convergence of HPC with AI.

Talk Abstract: Recent advancements in the field of Artificial Intelligence, particularly in the area of Deep Learning have left many traditional users of HPC somewhat unsure what benefits this might bring to their particular domain. What, for example, does the ability to identify members of felis catus from a selection of uploaded images on Facebook have to do with modeling the oceans of the world, or discovering how two molecules interact? This paper is designed to bridge the gap by identifying the state-of-the-art methodologies and use cases for applying AI to a range of computational science domains.

Session 2: The interplay of Compute / Data Technologies Unleashing A Digital Economy

Date and Time: December 19th, 2018: 10:00 AM – 12:00 Noon.

The theme for this year’s IRUS is the impact of Computational and Data Science technologies on emerging Digital Economy. The session will contain 3 speakers who will talk about their venture, their technology platform and approach from a business idea to a product/platform. The speakers will come from established as well as emerging players who are playing a key role in transforming the society such as democratizing financial sector using large scale citizen database, bringing affordable healthcare from remote diagnosis as well as insights from a technology vendor whose rapid advances often underpin such path-breaking applications.