Neural Network Verification with PyRAT

Augustin Lemesle, Julien Lehmann, Tristan Le Gall
10/31/2024
cs.AIcs.LG

No confident code implementation found for this paper.

Abstract

As AI systems are becoming more and more popular and used in various critical domains (health, transport, energy, ...), the need to provide guarantees and trust of their safety is undeniable. To this end, we present PyRAT, a tool based on abstract interpretation to verify the safety and the robustness of neural networks. In this paper, we describe the different abstractions used by PyRAT to find the reachable states of a neural network starting from its input as well as the main features of the tool to provide fast and accurate analysis of neural networks. PyRAT has already been used in several collaborations to ensure safety guarantees, with its second place at the VNN-Comp 2024 showcasing its performance.

View on arXivView PDF

Code Implementations

No confident code match yet

We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.

No code implementations found yet.

Know of an implementation? Let us know in the comments below!

Similar Papers

Verifying Safety of Neural Networks from Topological Perspectives

85% match2023

DeepCDCL: An CDCL-based Neural Network Verification Framework

84% match2024

SEEV: Synthesis with Efficient Exact Verification for ReLU Neural Barrier Functions

83% match1 repo32026

Automated Repair of AI Code with Large Language Models and Formal Verification

83% match2024

When to Trust AI: Advances and Challenges for Certification of Neural Networks

83% match2023

Formal Reasoning About Confidence and Automated Verification of Neural Networks

83% match2025

Cite this paper

@article{lemesle2024neural,
  title  = {Neural Network Verification with PyRAT},
  author = {Augustin Lemesle and Julien Lehmann and Tristan Le Gall},
  year   = {2024},
  eprint = {2410.23903},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2410.23903v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2410.23903.svg)](https://codeofpaper.com/paper/2410.23903)

Discussion