|
|
Kumar Kshitij Patel
Postdoctoral Associate, |
About Me
I am a postdoctoral associate at Yale FDS. I was previously a Simon's Research Fellow at the Simon's Institute for the Theory of Computing, UC Berkeley, for the Spring semester program on Federated and Collaborative Learning. Before joining Yale, I was a PhD student at the Toyota Technological Institute at Chicago (TTIC), where I had the privilege of being advised by Prof. Nati Srebro and Prof. Lingxiao Wang. Throughout my research career, I have explored various facets of collaborative learning, focusing on proving theoretical guarantees for optimization and ensuring the privacy of distributed algorithms amid data and systems heterogeneity. Recently, I have been interested in examining the incentives that encourage agents to initiate and sustain these collaborations (our recent workshop).
For an (mostly) up-to-date list of my publications, please visit my Google Scholar profile. You can also access my CV here.
During Summer 2023, I worked with Nidham Gazagnadou and Lingjuan Lyu from the Privacy Preserving Machine Learning team at Sony AI in Tokyo, Japan as a research intern. During summer 2020, I worked with the amazing team at Codeguru, Amazon Web Services as an applied scientist intern. And before joining TTIC, I obtained my BTech in Computer Science and Engineering at Indian Institute of Technology, Kanpur. There I was fortunate to work with Prof. Purushottam Kar on Bandit Learning algorithms. I also spent a year of my undergraduate on an academic exchange at École Polytechnique Fédérale de Lausanne (EPFL) where I worked at the Machine Learning and Optimization Laboratory (MLO) with Prof. Martin Jaggi.
News
[July 2026] I am giving a talk at the BSU seminar in the MRC Biostatistics Unit at the University of Cambridge. Due to some construction issues the seminar would be over zoom, but I am in Cambridge until July 25th and would love to meet!
[July 2026] I am visiting the UK From July 20th to August 2nd. Would love to meet and discuss research while I am here.
[July 2026] I gave an introductory lecture on differential privacy and training diffusion models with DP at Yale BDSY 2026. Here are the slides. Credits to Gautam Kamath's notes for how to introduce differential privacy to undergraduates.
[May 2026] I will be attending CVPR'26 in Denver to present our paper. If you wanna talk about research or anything else, ping me!
[April 2026] I will be attending ICLR'26 in Rio de Janeiro, Brazil. If you wanna talk about research or anything else, ping me!
[April 2026] I gave a spotlight talk about our ongoing work on data valuation at the Simons Institute's workshop on Agency in Collaborative Learning. Here is a video of the talk.
[March 2026] I am giving a talk on March 27th at the Machine Learning Seminar at UIUC. I will be in Urbana Champaign from March 26-29; let me know if you want to meet!
[March 2026] I am giving a virtual talk on March 23rd to the FL Group at Vector Institute.
[March 2026] I gave a talk on March 20th about our recent CVPR paper on differentially private federated training of personalized diffusion models at the Simon's workshop on Trust in Decentralized Systems. Here is a video of the talk.
[February 2026] Really excited to be co-organizing a Simon's workshop on data heterogeneity with some wonderful people (people close to me know that I spend more time than I would like to admit thinking about data heterogeneity assumptions :p).
[February 2026] Our paper on federated training of Diffusion models with personalization and Local Differential Privacy got accepted at CVPR'26. An updated version coming to Arxiv soon (here is an older version).
[January 2026] Our paper on group DRO Linear Regression got accepted at ICLR'26. An updated version coming to Arxiv soon (here is an older workshop version).
[January 2026] Excited to move to Berkeley for the semester-long research program on Federated and Collaborative Learning at the Simon's institute.
[December 2025] Coming to San Diego to present our new paper on Local SGD at NeurIPS'25.
[August 2025] Really excited to be co-organizing a TTIC summer workshop with some wonderful people.
[June 2025] I have graduated from TTIC and moved to Yale FDS as a postdoctoral associate. My thesis, titled "What Makes Local Updates Effective: The Role of Data Heterogeneity and Smoothness," can be found here.
Research Activities
I co-organized a workshop on Learning from Heterogeneous Sources as part of the Simon's Spring semester program on Federated and Collaborative Learning.
I co-organized a workshop on Incentives for Collaborative Learning and Data Sharing at TTIC this summer.
I co-organized a workshop on Theoritical Advances in Federated Learning last summer (2023) at TTIC.
I co-taught a tutorial at UAI'23 titled Online Optimization meets Federated Learning.
I served/am serving as a reviewer for STOC'21, TMLR, JMLR, ICML'21'22'24, NeurIPS'21'22'23'24, ICLR'22'23'24, AISTATS'22'23, Springer MLJ, as a session chair for ICML'22, NeurIPS'22, and as a volunteer for IJCAI'24, ICML'20, ICLR'20. I received the top reviewer award at ICLR'22, ICML'22, NeurIPS'22.
I am participating in the NSF-Simon's research collaboration on the Mathematics of Deep Learning (MoDL).
I co-organized the TTIC Student Workshop 2021, with Gene Li. We also organized a TTIC/Uchicago student theory seminar in Spring 2021. If you'd like to take over and re-start this series, please let me know.
I was a Teaching Assistant for the Convex Optmization course at TTIC during Winter'22'24 and a co-organizer for the Research at TTIC Colloquium for Fall-Winter 2021.
I participated in the Machine Learning Summer School at Tübingen, Germany during summer 2020.