Pinned Repositories
2d-shapley
This is an official repository for "2D-Shapley: A Framework for Fragmented Data Valuation" (ICML2023).
ASSET
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.
BEEAR
This is the official Gtihub repo for our paper: "BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models".
CLIP-MIA
This is an official repository for Practical Membership Inference Attacks Against Large-Scale Multi-Modal Models: A Pilot Study (ICCV2023).
LAVA
This is an official repository for "LAVA: Data Valuation without Pre-Specified Learning Algorithms" (ICLR2023).
Meta-Sift
The official implementation of USENIX Security'23 paper "Meta-Sift" -- Ten minutes or less to find a 1000-size or larger clean subset on poisoned dataset.
Narcissus
The official implementation of the CCS'23 paper, Narcissus clean-label backdoor attack -- only takes THREE images to poison a face recognition dataset in a clean-label way and achieves a 99.89% attack success rate.
privmon
This is an official repository for PrivMon: A Stream-Based System for Real-Time Privacy Attack Detection for Machine Learning Models (RAID 2023)
projektor
This is an official repository for "Performance Scaling via Optimal Transport: Enabling Data Selection from Partially Revealed Sources" (NeurIPS 2023).
Universal_Pert_Cert
This repo is the official implementation of the ICLR'23 paper "Towards Robustness Certification Against Universal Perturbations." We calculate the certified robustness against universal perturbations (UAP/ Backdoor) given a trained model.
ReDS Lab 's Repositories
reds-lab/Narcissus
The official implementation of the CCS'23 paper, Narcissus clean-label backdoor attack -- only takes THREE images to poison a face recognition dataset in a clean-label way and achieves a 99.89% attack success rate.
reds-lab/LAVA
This is an official repository for "LAVA: Data Valuation without Pre-Specified Learning Algorithms" (ICLR2023).
reds-lab/CLIP-MIA
This is an official repository for Practical Membership Inference Attacks Against Large-Scale Multi-Modal Models: A Pilot Study (ICCV2023).
reds-lab/Meta-Sift
The official implementation of USENIX Security'23 paper "Meta-Sift" -- Ten minutes or less to find a 1000-size or larger clean subset on poisoned dataset.
reds-lab/ASSET
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.
reds-lab/projektor
This is an official repository for "Performance Scaling via Optimal Transport: Enabling Data Selection from Partially Revealed Sources" (NeurIPS 2023).
reds-lab/Universal_Pert_Cert
This repo is the official implementation of the ICLR'23 paper "Towards Robustness Certification Against Universal Perturbations." We calculate the certified robustness against universal perturbations (UAP/ Backdoor) given a trained model.
reds-lab/BEEAR
This is the official Gtihub repo for our paper: "BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models".
reds-lab/2d-shapley
This is an official repository for "2D-Shapley: A Framework for Fragmented Data Valuation" (ICML2023).
reds-lab/privmon
This is an official repository for PrivMon: A Stream-Based System for Real-Time Privacy Attack Detection for Machine Learning Models (RAID 2023)
reds-lab/Forward-INF
reds-lab/Knowledge-Enriched-DMI
The official implementation of the ICCV 2021 paper, "Knowledge-Enriched Distributional Model Inversion Attacks."
reds-lab/Nash-Meta-Learning
Official implementation of "Fairness-Aware Meta-Learning via Nash Bargaining." We explore hypergradient conflicts in one-stage meta-learning and their impact on fairness. Our two-stage approach uses Nash bargaining to mitigate conflicts, enhancing fairness and model performance simultaneously.
reds-lab/I-BAU
Official Implementation of the ICLR 2022 paper, ``Adversarial Unlearning of Backdoors via Implicit Hypergradient''
reds-lab/Restricted_gradient_diversity_unlearning
reds-lab/Trojan_Removal_Benchmark
reds-lab/WokeyTalky
reds-lab/frequency-backdoor
The official implementation of the ICCV 2021 paper, "Rethinking the backdoor attacks' triggers: A frequency perspective."
reds-lab/preference-learning-with-rationales
This is the public repository for Data-Centric Human Preference Optimization with Rationales.
reds-lab/dataselection
Projektor Website
reds-lab/reds-lab.github.io
Homepage portfolio of Reds Projects
reds-lab/Woke-Pipeline