Title: Foundations of Robust Machine Learning under Imperfect Supervision

Speaker: Masashi Sugiyama, RIKEN/The University of Tokyo

Time and place: June 26, 15:00-16:00, SimulaMet, Stensberggata 27.

Online meeting link

Abstract:

In modern machine learning, acquiring vast amounts of high-quality labeled data is increasingly difficult and costly. While leveraging unlabeled data offers a potential alternative, it often lacks the necessary reliability for mission-critical applications. To bridge this gap, learning from "imperfect" supervision has emerged as a highly promising frontier. In this talk, I will review our recent advancements in developing robust frameworks for imperfect supervision, focusing on weakly supervised learning, noisy label learning, and transfer learning. I will conclude by discussing the strategic evolution of machine learning research in the transformative era of large foundation models.

Bio:

Masashi Sugiyama received his Ph.D. in Computer Science from Tokyo Institute of Technology, Japan, in 2001. After serving as an assistant and associate professor at the same institute, he became a professor at the University of Tokyo in 2014. Since 2016, he has also served as the director of the RIKEN Center for Advanced Intelligence Project.

His research interests include theories and algorithms of machine learning such as distribution shift adaptation, density ratio estimation, noise robust learning, weakly supervised learning, and
reinforcement learning. He served as a Program Co-Chair for ACML2010/2020, NeurIPS2015, and AISTATS2019.