ISIS Research Team Wins Best Paper Award at SmartComp 2019

The best paper at the 5th IEEE International conference on Smart Computing (IEEE Smartcomp 2019), held June 12-14 in Washington DC was titled “Mechanisms for Integrated Feature Normalization and Remaining Useful Life Estimation Using LSTMs Applied to Hard-Disks” and was authored by Sanchita Basak, Saptarshi Sengupta and Abhishek Dubey of Institute of Software Integrated Systems. LSTMs or Long Short-Term Memory Neural Networks are ideal for learning from temporal patterns. However, the mechanism fails when the data features are non-uniform. The paper improves state of the art by developing presenting statistical methods for data normalization and applying it to improve the prediction of remaining useful life. Prof. Dubey’s group have previously applied these techniques to batteries in small satellites. They are extending the work to be applied to engine telemetry for transit vehicles.

Sanchita Basak is a 2nd year Ph.D. student in the Department of EECS at Vanderbilt and her research focuses on designing data-driven neural network architectures for anomaly/fault detection and analyzing cascading effect of fault propagation in large-scale dynamical systems. Saptarshi Sengupta is a Ph.D. candidate in the Department of EECS at Vanderbilt and his research focuses on modeling and simulation of complex systems, multiagent systems and optimization under uncertainty. Dr. Abhishek Dubey is a Professor of Computer Science at Vanderbilt and his research is focused on Cyber Physical Systems, Resilient Distributed Systems and Fault Diagnosis.

The Smartcomp Best Paper Award recognizes the paper that exhibits key contributions in the field of smart computing. The best paper is nominated by the Smart comp organization committee. The ‘Best Paper’ award was presented to the authors by Dr. Wendy Nelson, the Program Director for the Smart and Connected Health Program in the Computer and Information Science and Engineering Directorate of the National Science Foundation.