Machine Learning Methods for Cayley Graphs Path Finding and Embeddings

Explore a 45-minute lecture from Institut des Hautes Etudes Scientifiques (IHES) that delves into applying machine learning and reinforcement learning techniques to analyze Cayley graphs, with particular emphasis on path finding and graph embeddings. Learn about an innovative approach inspired by DeepMind's AlphaGo system that surpasses traditional computer algebra systems like GAP, successfully finding paths in groups with orders of 10^40-10^70. Discover how this versatile method, applicable to any finite permutation or matrix group, produces shorter paths compared to conventional algorithmic solvers, including those designed for specific groups like the Rubik's Cube group. Understand how Cayley graphs serve as an excellent framework for comprehending key concepts in modern machine learning and reinforcement learning. Examine potential biological applications, including the construction of embeddings for proteins and small molecule drugs.