Universal scaling laws in the gradient descent training of neural networks
원문 초록
Current theoretical results on optimization trajectories of neural networks trained by gradient descent typically have the form of rigorous but potentially loose bounds on the loss values. In the present work we take a different approach and show that the learning trajectory can be characterized by an explicit asymptotic at large training times. Specifically, the leading term in the asymptotic expansion of the loss behaves as a power law $L(t) \sim t^{-\xi}$ with exponent $\xi$ expressed only through the data dimension, the smoothness of the activation function, and the class of function being approximated. Our results are based on spectral analysis of the integral operator representing the linearized evolution of a large network trained on the expected loss. Importantly, the techniques we employ do not require specific form of a data distribution, for example Gaussian, thus making our findings sufficiently universal.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Writing-RL: Advancing Long-form Writing via Adaptive Curriculum Reinforcement Learning
SuperWriter: Reflection-Driven Long-Form Generation with Large Language Models
How Does Response Length Affect Long-Form Factuality
MedScore: Factuality Evaluation of Free-Form Medical Answers
Frankentext: Stitching random text fragments into long-form narratives
Teaching Large Language Models to Maintain Contextual Faithfulness via Synthetic Tasks and Reinforcement Learning
토론
이 논문과 연결된 포럼 글이 아직 없습니다.