To verify and analyze classification properties of the neural networks when a small perturbation is applied to the image from MNIST, CIFAR-10 (and both with prepared Blured-Image) dataset.
-
Updated
Nov 20, 2021 - Python
#
deeppoly
Here are 6 public repositories matching this topic...
Formal verification tool for neural network robustness - critical for AI safety in high-stakes applications
-
Updated
Jun 6, 2025 - Python
This is the course project for Reliable and Trustworthy Artificial Intelligence course 2022 given by Prof. Martin Vechev. at ETH Zurich.
-
Updated
Dec 14, 2022 - Python
Neural network verifier using DeepPoly with learnable ReLU relaxation — ETH Zurich RIAI 2024
pytorch eth-zurich abstract-interpretation adversarial-robustness neural-network-verification deeppoly
-
Updated
Feb 24, 2026 - Python
Sound neural-network robustness verifier based on DeepPoly abstract interpretation, back-substitution + learnable ReLU (α) bounds certify L∞ adversarial robustness on MNIST/CIFAR-10 classifiers.
deep-learning pytorch mnist formal-verification abstract-interpretation cifar10 adversarial-examples adversarial-robustness robustness-verification neural-network-verification certified-robustness convex-relaxation deeppoly
-
Updated
Jun 4, 2026 - Python
Improve this page
Add a description, image, and links to the deeppoly topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the deeppoly topic, visit your repo's landing page and select "manage topics."