Towards Analytic Learning Theory: Regret Analysis

Explore a 22-minute conference talk by Wojciech Szpankowski from Purdue University, delivered at the International Conference on Probabilistic, Combinatorial and Asymptotic Methods for the Analysis of Algorithms (AofA 2025) on May 6, 2025. Delve into the emerging field of analytic learning theory with a focus on regret analysis, which applies analytic combinatorics and probabilistic methods to understand fundamental learning algorithms. Gain insights into how mathematical analysis can be used to evaluate and improve machine learning performance through rigorous regret bounds.