Structured ML Training via Conditional Gradients
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Explore a 24-minute conference talk on structured machine learning training using conditional gradients. Delve into Sebastian Pokutta's presentation from the Deep Learning and Combinatorial Optimization 2021 event at the Institute for Pure & Applied Mathematics (IPAM). Learn about the importance of conditional gradient methods in minimizing convex functions over polytopes, their applications in structured optimization and learning, and recent developments in both traditional optimization and deep learning. Gain insights into how these methods incorporate polyhedral structure into solutions, presented by an expert from the Konrad-Zuse-Zentrum für Informationstechnik (ZIB) Department of Mathematics.
Syllabus
Sebastian Pokutta: "Structured ML Training via Conditional Gradients"
Taught by
Institute for Pure & Applied Mathematics (IPAM)