2 papers
cs.AI2026
Grokking on the Weight-Decay Clock: A Rate Hierarchy from Softly Broken Symmetries
Taeyoung Kim
Delayed generalization, or grokking, remains poorly understood despite extensive empirical study. We identify an exactly solvable late-time relaxation mechanism for grokking in lin…
cs.LG2024
An Infinite-Width Analysis on the Jacobian-Regularised Training of a Neural Network
Taeyoung Kim, Hongseok Yang
The recent theoretical analysis of deep neural networks in their infinite-width limits has deepened our understanding of initialisation, feature learning, and training of those net…