BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/1290
DTSTAMP:20230914T125958Z
SUMMARY:Towards Next-Generation ML/AI: Robustness\, Optimization\, Privacy
DESCRIPTION:Speaker: Krishna Pillutla (Google Research\, U.S.A.)\n\nAbstrac
 t: \nTwo trends have taken hold in machine learning and artificial intelli
 gence: a move to massive\, general-purpose\, pre-trained models as well as
  a move to small\, on-device models trained on distributed data. Both thes
 e disparate settings face some common challenges: a need for (a) robustnes
 s to deployment conditions that differ from training\, (b) faster optimiza
 tion\, and (c) protection of data privacy.\nAs a result of the former tren
 d\, large language models have displayed emergent capabilities they have n
 ot been trained for. Recent models such as ChatGPT have attained the abili
 ty to generate remarkably human-like long-form text. I will describe Mauve
 \, a measure to quantify this ability by measuring the gap between the dis
 tribution of generated text and that of human-written text. I will highlig
 ht its good empirical performance and present some statistical estimation 
 results.\nThe move to massively distributed on-device federated learning o
 f models opens up new challenges due to the natural diversity of the under
 lying user data and the need to protect its privacy. I will discuss how to
  reframe the learning problem to make the model robust to natural distribu
 tion shifts arising from deployment on diverse users who do not conform to
  the population trends in a manner that admits a distributed optimization 
 algorithm with end-to-end differential privacy.\nTo conclude\, I will disc
 uss my ongoing efforts and future plans to work toward the next generation
  of ML/AI techniques by combining the best of both worlds. I will discuss 
 applications ranging from differentially private language models and text 
 generation to decentralized learning.\n\nBio: Krishna Pillutla is a visiti
 ng researcher (postdoc) at Google Research\, USA in the Federated Learning
  team. He obtained his Ph.D. at the University of Washington where he was 
 advised by Zaid Harchaoui and Sham Kakade. Before that\, he received his M
 .S. from Carnegie Mellon University and B.Tech. from IIT Bombay where he w
 as advised by Nina Balcan and J. Saketha Nath respectively. Krishna's rese
 arch has been recognized by a NeurIPS outstanding paper award (2021) and a
  JP Morgan Ph.D. fellowship (2019-20).\n
URL:https://www.tcs.tifr.res.in/web/events/1290
DTSTART;TZID=Asia/Kolkata:20230502T190000
DTEND;TZID=Asia/Kolkata:20230502T200000
LOCATION:Online
END:VEVENT
END:VCALENDAR