BNN verification dataset for Max-SAT Evaluation 2020

A tic-tac-toe using minmax adversial search algorithm as an opponent

LittleAdversary is an adversarial machine learning library made to aid research into adversarial attacks and defences, with a primary focus on one-shot defences. It contains an end-to-end implementation of the proposed defence in 'Siamese Neural Networks for Adversarial Robustness ', complete with statistical analysis of the results.

http://excel.fit.vutbr.cz/submissions/2020/029/29.pdf

The impact of transitive annotation on the training of taxonomic classifiers

Crafting adversarial examples with one pixel attack

Code for paper Geometric properties of adversarial images [DSMP 2020]

终于有自己的博客啦~记录自己的学习心得，陆续搬运上传ing

Framework for generating Adversarial Attacks on Deep Neural Networks using Evolutionary Strategies (ES).

Recognition by Components

Source of the ECCV22 paper "LGV: Boosting Adversarial Example Transferability from Large Geometric Vicinity"

An unofficial version of the PyTorch implementation of CURE and Fast Adversarial training with FGSM.

[UAI 2024 paper] DistriBlock: Identifying adversarial audio samples by leveraging characteristics of the output distribution.

Defending Neural Networks from Adversarial Attacks

Framework for creating Adversarial Attacks on Deep Neural Networks with Evolutionary Strategies (ES).

Review and analysis of selected adversarial attacks. We implement common attack methods and evaluate them with a GoogleNet network on ImageNet like data.

Explanation-guided boosting of machine learning evasion attacks.

B.Sc. Final Project: Generating adversarial examples using GAN (Generative Adversarial Network) in Pytorch on the MNIST dataset.

Training robust models takes a looong time since generating adversaries is so expensive. We design a parallel algorithm for training large robust models using PyTorch C++/MPI and show it runs fast!