MaLGa Colloquia - The Statistical Complexity of Interactive Decision Making
Alexander (Sasha) Rakhlin - Massachusetts Institute of Technology
In this talk, we will present a new emerging framework for interactive decision making. We will examine multi-armed bandits, contextual bandits, structured bandits, general decision making, and reinforcement learning under the same umbrella. We will present (i) sample complexity measures for interactive decision making that are necessary and sufficient for sample-efficient learning, and (ii) unified algorithm design principles that achieve optimal sample complexity.
Alexander (Sasha) Rakhlin is a Professor at MIT, with appointments in the Statistics & Data Science Center and the Department of Brain and Cognitive Sciences. Prior to joining MIT, Sasha was with the Department of Statistics at the University of Pennsylvania. His interests are in mathematical statistics and machine learning.
May 22nd 2023, 16:00
Room 704, UniGe DIMA, Via Dodecaneso 35