Rajeev Rastogi

Vice President, Machine Learning at Amazon

Bengaluru, India

Topic:  CRISP: A Probabilistic Model for Individual-Level COVID-19 Infection Risk Estimation Based on Contact Data

Rajeev Ramnarain Rastogi is an Indian computer scientist who graduated from the Indian Institute of Technology Bombay. He received his Master’s and Doctoral degrees from the University of Texas in 1990 and 1993 respectively. In 1993 Rastogi started working at Bell Labs in Murray Hill, New Jersey. He became member of technical staff at its Information Sciences Research Center. Five years later he held the Distinguished Member of Technical Staff position and by 1999 became a director of the Internet Management Research Department and became a Bell Labs fellow in 2003. In 2012, he became a fellow of the Association for Computing Machinery “for contributions to the analysis and management of large data sets.” He was Vice President of Yahoo! Labs in Bangalore  and is currently serving as a director of Machine Learning on Amazon.com. He has over 200 peer-reviewed articles with the CURE: An Efficient Clustering Algorithm for Large Databases which received over 3,100 citations since 1998, bringing him an h-index of 63.

Rajeev is recipient of Bell Labs Fellow, ACM Fellow and IEEE ICDM Research Contributions Award.

Abstract:

In this talk, I will present CRISP, a probabilistic graphical model for COVID-19 infection spread
through a population based on the SEIR model where we assume access to (1) mutual contacts
between pairs of individuals across time across various channels (e.g., Bluetooth contact traces),
as well as (2) test outcomes at given times for infection. Our micro-level model keeps track of
the infection state for each individual at every point in time, ranging from susceptible, exposed,
infectious to recovered. We develop a Monte Carlo EM algorithm to infer contact-channel
specific infection transmission probabilities. Our algorithm uses Gibbs sampling to draw samples
of the latent infection status of each individual over the entire time period of analysis, given the
latent infection status of all contacts and test outcome data. Experimental results with simulated
data demonstrate that a testing-and-quarantining policy based on infection risk scores computed
by the CRISP algorithm is able to effectively mitigate COVID-19 infection spread.