A Generalization Bound for Nearly-Linear Networks

Autor:	Golikov, Eugene
Rok vydání:	2024
Předmět:	Computer Science - Machine Learning Computer Science - Artificial Intelligence Statistics - Machine Learning
Druh dokumentu:	Working Paper
Popis:	We consider nonlinear networks as perturbations of linear ones. Based on this approach, we present novel generalization bounds that become non-vacuous for networks that are close to being linear. The main advantage over the previous works which propose non-vacuous generalization bounds is that our bounds are a-priori: performing the actual training is not required for evaluating the bounds. To the best of our knowledge, they are the first non-vacuous generalization bounds for neural nets possessing this property. Comment: 22 pages, 9 figures
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2407.06765 Zobrazit plný text záznamu View this record from Arxiv