Sample Compression and Topological Radon Theorem in Machine Learning

Explore a fascinating lecture that bridges topology and machine learning, delving into the application of topological Radon theorem to solve theoretical machine learning challenges, particularly focusing on sample compression schemes. Learn how mathematical tools beyond traditional statistics, probability, and combinatorics can be effectively applied to machine learning problems through this 56-minute presentation delivered at HUJI Machine Learning Club. Discover insights from postdoctoral researcher Bogdan Chornomaz's collaborative work with Zachary Chase, Steve Hanneke, Shay Moran, and Amir Yehudayoff, examining how seemingly exotic mathematical concepts find natural applications in machine learning theory. Gain understanding of the intersection between topology and machine learning through the lens of sample compression and the topological Radon theorem's variant applications.