Showing cs.LGShow all
2 papers · 1 filter
cs.LG2025
Globally aware optimization with resurgence
Wei Bu
Modern optimization faces a fundamental challenge: local gradient-based methods provide no global information about the objective function landscape, often leading to suboptima…
cs.LG2025
Fokker-Planck to Callan-Symanzik: evolution of weight matrices under training
Wei Bu, Uri Kol, Ziming Liu
The dynamical evolution of a neural network during training has been an incredibly fascinating subject of study. First principal derivation of generic evolution of variables in sta…