Provable Generalization of SGD-Trained Neural Networks of Any Width in the Presence of Adversarial Label Noise
Abstract
We consider a one-hidden-layer leaky ReLU network of arbitrary width trained by stochastic gradient descent (SGD) following an arbitrary initialization. We prove that SGD produces neural networks that have classification accuracy competitive with that of the best halfspace over the distribution for a broad class of distributions that includes log-concave isotropic and hard margin distributions. Equivalently, such networks can generalize when the data distribution is linearly separable but corrupted with adversarial label noise, despite the capacity to overfit. To the best of our knowledge, this is the first work to show that overparameterized neural networks trained by SGD can generalize when the data is corrupted with adversarial label noise.
Cite
Text
Frei et al. "Provable Generalization of SGD-Trained Neural Networks of Any Width in the Presence of Adversarial Label Noise." International Conference on Machine Learning, 2021.Markdown
[Frei et al. "Provable Generalization of SGD-Trained Neural Networks of Any Width in the Presence of Adversarial Label Noise." International Conference on Machine Learning, 2021.](https://mlanthology.org/icml/2021/frei2021icml-provable/)BibTeX
@inproceedings{frei2021icml-provable,
title = {{Provable Generalization of SGD-Trained Neural Networks of Any Width in the Presence of Adversarial Label Noise}},
author = {Frei, Spencer and Cao, Yuan and Gu, Quanquan},
booktitle = {International Conference on Machine Learning},
year = {2021},
pages = {3427-3438},
volume = {139},
url = {https://mlanthology.org/icml/2021/frei2021icml-provable/}
}