Quick Links
Sayak Ray Chowdhury
PhD (Indian Institute of Science, Bangalore)
Assistant Professor, Department of Computer Science & Engineering
Research Interest
Machine Learning, Artificial Intelligence, Differential Privacy
KD 223, H.R. Kadim Diwan Building
Specialization
Sequential Decision Making, Learning Theory, Privacy-Preserving Machine Learning, Language Model Optimization
Education
PhD (2021): Indian Institute of Science, Bangalore
M.E. (2015): Indian Institute of Science, Bangalore
B.E. (2012): Jadavpur University, Kolkata
Previous Work Experience
Postdoctoral Researcher (2022 - 2024): Microsoft Research, India
Postdoctoral Associate (2021 - 2022): Boston University, USA
Selected Publications
Communication Efficient Secure and Private Multi-Party Deep Learning. Sankha Das, Sayak Ray Chowdhury, Nishanth Chandran, Divya Gupta, Rahul Sharma, Satya Lokam. Proceedings on Privacy Enhancing Technologies Symposium (PETS), 2025.
Distributed Differential Privacy in Multi-armed Bandits. Sayak Ray Chowdhury, Xingyu Zhou. International Conference on Learning Representations (ICLR), 2023.
Provably Robust DPO: Aligning Language Models with Noisy Feedback. Sayak Ray Chowdhury, Anush Kini, Nagarajan Natarajan. International Conference on Machine Learning (ICML), 2024.
Differentially Private Federated Linear Contextual Bandits. Xingyu Zhou, Sayak Ray Chowdhury. International Conference on Learning Representations (ICLR), 2024.
Differentially Private Reward Estimation with Preference Feedback. Sayak Ray Chowdhury, Xingyu Zhou, Nagarajan Natarajan. International Conference on Artificial Intelligence and Statistics (AISTATS), 2024.
Bregman Deviations of Generic Exponential Families. Sayak Ray Chowdhury, Patrick Saux, Odalric-Ambrym Maillard, Aditya Gopalan. 36th Annual Conference on Learning Theory (COLT), 2023. Link
Shuffle Private Linear Contextual Bandits. Sayak Ray Chowdhury, Xingyu Zhou. International Conference on Machine Learning (ICML), 2022.
Bayesian Optimization under Heavy-tailed Payoffs. Sayak Ray Chowdhury, Aditya Gopalan. Neural Information Processing Systems (NeurIPS), 2019
Online Learning in Kernelized Markov Decision Processes. Sayak Ray Chowdhury, Aditya Gopalan. International Conference on Artificial Intelligence and Statistics (AISTATS), 2019.
On Kernelized Multi-armed Bandits. Sayak Ray Chowdhury, Aditya Gopalan. International Conference on Machine Learning (ICML), 2017.
Awards & Fellowships
Google Research Ph.D. Fellowship (2017-2021)
Division of Systems Engineering, Boston University Postdoctoral Fellowship (2021)
Research Group
Machine Intelligence, Learning, Automation, and Privacy (MILAP)
Webpage: https://sites.google.com/view/sayakraychowdhury/group
UG/PG Courses Developed
CS 774: Differential Privacy in Machine Learning
Invited Lectures
June 2024: Provably Robust DPO: Aligning Language Models with Noisy Preferences. TrustML Young Scientists Seminar, RIKEN & University of Tokyo, Japan.
July 2024: Differential Privacy in Learning from Preference Data. Bangalore Crypto Day, IISc Bangalore, India.
June 2023: Differential Privacy in Reinforcement Learning. Machine Learning Summer School for Science, Jagiellonian University, Poland.
November 2022: Time Uniform Concentration Bounds for Black Box Optimization. IEEE Information Theory Workshop (ITW), IIT Bombay, India.