Efficient Statistical Assessment of Neural Network Corruption Robustness

Abstract

We quantify the robustness of a trained network to input uncertainties with a stochastic simulation inspired by the field of Statistical Reliability Engineering. The robustness assessment is cast as a statistical hypothesis test: the network is deemed as locally robust if the estimated probability of failure is lower than a critical level.The procedure is based on an Importance Splitting simulation generating samples of rare events. We derive theoretical guarantees that are non-asymptotic w.r.t. sample size. Experiments tackling large scale networks outline the efficiency of our method making a low number of calls to the network function.

Cite

Text

Tit et al. "Efficient Statistical Assessment of Neural Network Corruption Robustness." Neural Information Processing Systems, 2021.

Markdown

[Tit et al. "Efficient Statistical Assessment of Neural Network Corruption Robustness." Neural Information Processing Systems, 2021.](https://mlanthology.org/neurips/2021/tit2021neurips-efficient/)

BibTeX

@inproceedings{tit2021neurips-efficient,
  title     = {{Efficient Statistical Assessment of Neural Network Corruption Robustness}},
  author    = {Tit, Karim and Furon, Teddy and Rousset, Mathias},
  booktitle = {Neural Information Processing Systems},
  year      = {2021},
  url       = {https://mlanthology.org/neurips/2021/tit2021neurips-efficient/}
}