adversarial-robustness 开源项目推荐（294 个）

RobustBench /robustbench

RobustBench：标准化的对抗鲁棒性基准 [NeurIPS 2021 基准与数据集赛道]

fra31 /auto-attack

与“Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks”相关的代码

thu-ml /ares

一个专注于对抗鲁棒性基准测试的对抗机器学习 Python 库。

Verified-Intelligence /alpha-beta-CROWN

alpha-beta-CROWN：一个高效、可扩展且 GPU 加速的神经网络验证器（VNN-COMP 2021、2022、2023、2024、2025 冠军）

alibaba /easyrobust

EasyRobust：一个易用的库，用于基于 PyTorch 开展最先进的鲁棒计算机视觉研究。

max-andr /square-attack

Square Attack：一种通过随机搜索实现的查询高效黑盒对抗攻击 [ECCV 2020]

x-zheng16 /Awesome-Embodied-AI-Safety

Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses | 500+ Papers | Perception, Cognition, Planning, Interaction, Agentic System

LayneH /self-adaptive-training

[TPAMI2022 & NeurIPS2020] Official implementation of Self-Adaptive Training

VITA-Group /Aug-NeRF

[CVPR 2022] "Aug-NeRF: Training Stronger Neural Radiance Fields with Triple-Level Physically-Grounded Augmentations" by Tianlong Chen*, Peihao Wang*, Zhiwen Fan, Zhangyang Wang

microsoft /denoised-smoothing

可证明地防御预训练分类器，包括 Azure、Google、AWS 和 Clarifai API

imrahulr /adversarial_robustness_pytorch

Unofficial implementation of the DeepMind papers "Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examples" & "Fixing Data Augmentation to Improve Adversarial Robustness" in PyTorch

AI-secure /InfoBERT

[ICLR 2021]“InfoBERT：从信息论角度提高语言模型的鲁棒性”作者：王博欣、王硕航、宇成、甘哲、贾若曦、李波、刘婷婷

VITA-Group /Adv-SS-Pretraining

[CVPR 2020] Adversarial Robustness: From Self-Supervised Pre-Training to Fine-Tuning

Haichao-Zhang /FeatureScatter

Feature Scattering Adversarial Training (NeurIPS19)

adversarial-for-goodness /Co-Attack

[ACM MM 2022] Official PyTorch implementation of Towards Adversarial Attack on Vision-Language Pre-training Models (Co-Attack)

hxxdtd /Awesome-Diffusion-Model-Unlearning

A repository of resources on machine unlearning for diffusion models

cemanil /LNets

Lipschitz Neural Networks described in "Sorting Out Lipschitz Function Approximation" (ICML 2019).

zjysteven /DVERGE

[NeurIPS'20 Oral] DVERGE: Diversifying Vulnerabilities for Enhanced Robust Generation of Ensembles

jiequancui /DKL

Decoupled Kullback-Leibler Divergence Loss (DKL), NeurIPS 2024 / Generalized Kullback-Leibler Divergence Loss (GKL), TPAMI 2026

VITA-Group /Alleviate-Robust-Overfitting

[ICLR 2021] "Robust Overfitting may be mitigated by properly learned smoothening" by Tianlong Chen*, Zhenyu Zhang*, Sijia Liu, Shiyu Chang, Zhangyang Wang