Vinith Suriyakumar

University of Toronto, Vector Institute, The Hospital for Sick Children

I’m a Masters student at the University of Toronto & Vector Institute advised by Dr. Marzyeh Ghassemi, Dr. Nicolas Papernot, Dr. Berk Ustun, and Dr. Anna Goldenberg. My research focuses on fundamental questions of trustworthy machine learning and responsible computing. Currently, I focus on the theory and applications of differential privacy, algorithmic fairness, and cryptography to machine learning and healthcare. I use techniques from math, statistics, optimization, economics, and theoretical computer science (i.e. learning theory and computational complexity) to conduct this research.

Current Projects

  1. Differentially Private Stochastic Adversarial Optimization
  2. Fair Personalization in Machine Learning for Healthcare
  3. Privacy Amplification for Sequential & Correlated Data

news

Mar 29, 2021 I’m excited to announce that I’ve accepted an offer to join the PhD program starting Fall 2021 at MIT EECS where I will be affiliated with CSAIL, LIDS, and IMES. I will be co-advised by Dr. Ashia Wilson and Dr. Marzyeh Ghassemi. I will continue my research focusing on the methodological and theoretical foundations of machine learning, optimization, sampling, differential privacy, and algorithmic fairness.
Feb 19, 2021 I’m excited to announce that I’ve accepted an offer to join Google this summer as a Research Intern working with Dr. Om Thakkar on differentially private federated learning!
Jan 4, 2021 I’ll be giving a long talk on the “Challenges of Differentially Private Prediction in Health Care” at the IJCAI 2021 AI for Social Good Workshop organized by Harvard CRCS on January 7th during the 8:10-9:10 PM EST session. A longer version of this study was recently accepted for publication at ACM FAccT 2021.
Dec 17, 2020 I had two papers accepted at ACM FAccT 2021: Chasing Your Long Tails: Differentially Private Prediction in Health Care Settings and “Can You Fake It Until You Make It?: Impacts of Differentially Private Synthetic Data on Downstream Classification Fairness” (preprint forthcoming)!
Oct 15, 2020 Our preprint Chasing Your Long Tails: Differentially Private Prediction in Health Care Settings is up! I’ll be giving a talk virtually on our work at the University of Toronto, Centre for Ethics on October 28, 2020 from 4:00 - 5:00 PM EST.