Prithwish Chakraborty, Ph.D.
The ready availability of Electronic Health Record (EHR) systems has allowed providers and researchers to obtain a longitudinal view of disease and treatment progression from patient encounter histories. These datasets can cover various facets of patient medical history such as diagnoses, prescribed medications, and laboratory observations. Intelligent analysis of such meta-records can lead to an improved understanding of diseases, outcomes, treatments, or even lead to a revision of recommended treatment plans. While classical analyses of these records have shown the importance of such datasets, machine learning (classical and deep learning) approaches are increasingly being applied to support these analyses. This talk will explore three specific EHR problems viz (i) Computational Phenotyping, (ii) Disease Progression Modeling, and (iii) Risk estimation in diseases. The talk will discuss some classical approaches followed by deeper dives into a few of our contributions within IBM Center for Computational Health in each of these problems.
Prithwish Chakraborty is currently a Research Staff Member in the Center for Computational Health at the IBM T.J. Watson Research Center, NY. His work focuses on applications of data science towards patient health characterization and risk modeling. Broadly, his research interests are temporal data mining, machine learning and image recognition. He completed his Ph.D. in computer science from Virginia Tech in 2016. He has received manager choice and excellence awards at IBM. He led a team from IBM Research that won the first biobank disease challenge organized by Partners Healthcare in 2018. He has published ~30 papers with more than 750 citations to premier journals and conferences such as KDD, SDM, AAAI, and TBME. He co-organized DSHealth 2019 workshop and MLMH 2018 workshop as part of the Knowledge Discovery and Data Mining conference, KDD 2018 and KDD 2019 respectively. He serves on the program committee and regular reviewer for venues such as IJCAI, AAAI, NeurIPS, ICML, and TNNLS.
To RSVP, contact Yue Ning at [email protected]