A list of backdoor learning resources

A curated list of papers & resources on backdoor attacks and defenses in deep learning.

A curated list of papers & resources linked to data poisoning, backdoor attacks and defenses against them

This is an implementation demo of the ICLR 2021 paper [Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks](https://openreview.net/pdf?id=9l0K4OM-oXE) in PyTorch.

BackdoorSim: An Educational into Remote Administration Tools

Codes for NeurIPS 2021 paper "Adversarial Neuron Pruning Purifies Backdoored Deep Models"

Official Implementation of ICLR 2022 paper, ``Adversarial Unlearning of Backdoors via Implicit Hypergradient''

FLIP: A Provable Defense Framework for Backdoor Mitigation in Federated Learning [ICLR‘23, Best Paper Award at ECCV’22 AROW Workshop]

[ICLR2023] Distilling Cognitive Backdoor Patterns within an Image

This is an implementation demo of the IJCAI 2022 paper [Eliminating Backdoor Triggers for Deep Neural Networks Using Attention Relation Graph Distillation](https://arxiv.org/abs/2204.09975) in PyTorch.

This repository is the official implementation of the paper "ASSET: Robust Backdoor Data Detection Across a Multiplicity of Deep Learning Paradigms." ASSET achieves state-of-the-art reliability in detecting poisoned samples in end-to-end supervised learning/ self-supervised learning/ transfer learning.

[Findings of EMNLP 2022] Expose Backdoors on the Way: A Feature-Based Efficient Defense against Textual Backdoor Attacks

Implementation of "Beating Backdoor Attack at Its Own Game" (ICCV-23).

Backdoor Stuff in AI/ ML domain

🪨️🚪️📖️ The official documentation source repository for the Doorcrete project.

Mask and Restore: Blind Backdoor Defense at Test Time with Masked Autoencoder

Implementation of ICCV'23 paper on Multimodal Backdoor Defense Technique: TIJO (Trigger Inversion using Joint Optimization)

This is a project used for detecting backdoors at different levels.

Monitoring Your User Uploader Activity

AI/ ML papers in DBLP/ arXiv