auto_LiRPA: An Automatic Linear Relaxation based Perturbation Analysis Library for Neural Networks and General Computational Graphs
-
Updated
Apr 13, 2024 - Python
auto_LiRPA: An Automatic Linear Relaxation based Perturbation Analysis Library for Neural Networks and General Computational Graphs
Neural Network Verification Software Tool
alpha-beta-CROWN: An Efficient, Scalable and GPU Accelerated Neural Network Verifier (winner of VNN-COMP 2021, 2022, and 2023)
Formal Verification of Neural Feedback Loops (NFLs)
Certified defense to adversarial examples using CROWN and IBP. Also includes GPU implementation of CROWN verification algorithm (in PyTorch).
Reference implementations for RecurJac, CROWN, FastLin and FastLip (Neural Network verification and robustness certification algorithms) [Do not use this repo, use https://github.com/Verified-Intelligence/auto_LiRPA instead]
[NeurIPS 2019] H. Chen*, H. Zhang*, S. Si, Y. Li, D. Boning and C.-J. Hsieh, Robustness Verification of Tree-based Models (*equal contribution)
β-CROWN: Efficient Bound Propagation with Per-neuron Split Constraints for Neural Network Verification
[NeurIPS 2021] Towards Better Understanding of Training Certifiably Robust Models against Adversarial Examples | ⛰️
[ICLR 2020] Code for paper "Robustness Verification for Transformers"
[CCS 2021] TSS: Transformation-specific smoothing for robustness certification
Fast Adversarial Robustness Certification of Nearest Prototype Classifiers for Arbitrary Seminorms [NeurIPS 2020]
Sampling-based Scalable Quantitative Verification for DNNs
This github repository contains the official code for the paper, "Evolving Robust Neural Architectures to Defend from Adversarial Attacks"
This code base is intended to serve as a starting point for interested researchers or practitioners to extend or apply the robustness verification portion of the author's Master's thesis " GUM-compliant neural-network robustness verification".
The official repo for GCP-CROWN paper
Benchmark for formally verifying ViTs
certifying robustness of neural network via convex optimization
An algorithm to calculate the convex hull of ReLU function for neural network verification.
DPLL(T)-based Verification tool for DNNs
Add a description, image, and links to the robustness-verification topic page so that developers can more easily learn about it.
To associate your repository with the robustness-verification topic, visit your repo's landing page and select "manage topics."