On Lipschitz Regularization of Convolutional Layers Using Toeplitz Matrix Theory
Abstract
This paper tackles the problem of Lipschitz regularization of Convolutional Neural Networks. Lipschitz regularity is now established as a key property of modern deep learning with implications in training stability, generalization, robustness against adversarial examples, etc. However, computing the exact value of the Lipschitz constant of a neural network is known to be NP-hard. Recent attempts from the literature introduce upper bounds to approximate this constant that are either efficient but loose or accurate but computationally expensive. In this work, by leveraging the theory of Toeplitz matrices, we introduce a new upper bound for convolutional layers that is both tight and easy to compute. Based on this result we devise an algorithm to train Lipschitz regularized Convolutional Neural Networks.
Cite
Text
Araujo et al. "On Lipschitz Regularization of Convolutional Layers Using Toeplitz Matrix Theory." AAAI Conference on Artificial Intelligence, 2021. doi:10.1609/AAAI.V35I8.16824Markdown
[Araujo et al. "On Lipschitz Regularization of Convolutional Layers Using Toeplitz Matrix Theory." AAAI Conference on Artificial Intelligence, 2021.](https://mlanthology.org/aaai/2021/araujo2021aaai-lipschitz/) doi:10.1609/AAAI.V35I8.16824BibTeX
@inproceedings{araujo2021aaai-lipschitz,
title = {{On Lipschitz Regularization of Convolutional Layers Using Toeplitz Matrix Theory}},
author = {Araujo, Alexandre and Négrevergne, Benjamin and Chevaleyre, Yann and Atif, Jamal},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2021},
pages = {6661-6669},
doi = {10.1609/AAAI.V35I8.16824},
url = {https://mlanthology.org/aaai/2021/araujo2021aaai-lipschitz/}
}