A Generalization Bound for Nearly-Linear Networks

arXiv:2407.06765v2 Announce Type: replace Abstract: 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.

Source: arXiv cs.LG | 2026-06-23