Syomantak Chaudhuri 

I am a fifth-year Ph.D. student at the EECS department, UC Berkeley. I am fortunate to be advised by Tom Courtade.

My Ph.D. research mainly focuses on dealing with heterogeneity of samples in statistical tasks.  Some aspects of heterogeneity that we have worked on include questions like

1. what are the optimal mechanisms and corresponding minimax rates for estimating mean from user data where each user has a different privacy demand?

2. how can we provide different levels of privacy to different features (or coordinates) of a user's data?

3. what are the optimal algorithms and corresponding minimax rates for robust mean estimation and linear regression subject to different samples being corrupted with different probabilities?

I have interned at Google Research India, Microsoft USA, and Susquehanna International Group USA.