paper-with-me

Papers

PureEBM: Universal Poison Purification via Mid-Run Dynamics of Energy-Based Models

2024-05-28 · Omead Pooladzandi, Jeffrey Jiang, Sunay Bhat, Gregory Pottie

Data poisoning attacks pose a significant threat to the integrity of machine learning models by leading to misclassification of target distribution data by injecting adversarial examples during training. Existing state-of-the-art (SoTA) defense methods suffer from limitations, such as significantly reduced generalization performance and significant overhead during training, making them impractical or limited for real-world applications. In response to this challenge, we introduce a universal data purification method that defends naturally trained classifiers from malicious white-, gray-, and black-box image poisons by applying a universal stochastic preprocessing step $\Psi_{T}(x)$, realized by iterative Langevin sampling of a convergent Energy Based Model (EBM) initialized with an image $x.$ Mid-run dynamics of $\Psi_{T}(x)$ purify poison information with minimal impact on features important to the generalization of a classifier network. We show that EBMs remain universal purifiers, even in the presence of poisoned EBM training data, and achieve SoTA defense on leading triggered and triggerless poisons. This work is a subset of a larger framework introduced in \pgen with a more detailed focus on EBM purification and poison defense.

📄 PDF Abstract BibTeX arXiv:2405.19376

Code (1)

SunayBhat1/PureGen_PoisonDefense 공식 구현 pytorch

Tasks

Data Poisoning

Methods 이 논문이 사용한 방법론

Focus 설명 없음
EBM 설명 없음

Similar Papers 제목 키워드 기반

PureGen: Universal Data Purification for Train-Time Poison Defense via Generative Model Dynamics

2024-05-28 · Sunay Bhat, Jeffrey Jiang, Omead Pooladzandi, Alexander Branch 외

Train-time data poisoning attacks threaten machine learning models by introducing adversarial examples during training, leading to misclassification. Current defense methods often reduce generalization performance, are a…

Data PoisoningDenoising

ECLIPSE: Expunging Clean-label Indiscriminate Poisons via Sparse Diffusion Purification

2024-06-21 · Xianlong Wang, Shengshan Hu, Yechao Zhang, Ziqi Zhou 외

Clean-label indiscriminate poisoning attacks add invisible perturbations to correctly labeled training images, thus dramatically reducing the generalization capability of the victim models. Recently, some defense mechani…

Denoising

FLARE: Toward Universal Dataset Purification against Backdoor Attacks

2024-11-29 · Linshan Hou, Wei Luo, Zhongyun Hua, Songhua Chen 외

Deep neural networks (DNNs) are susceptible to backdoor attacks, where adversaries poison datasets with adversary-specified triggers to implant hidden backdoors, enabling malicious manipulation of model predictions. Data…

All

DUP: Detection-guided Unlearning for Backdoor Purification in Language Models

2025-08-03 · Man Hu, Yahui Ding, Yatao Yang, Liangyu Chen 외 arxiv

As backdoor attacks become more stealthy and robust, they reveal critical weaknesses in current defense strategies: detection methods often rely on coarse-grained feature statistics, and purification methods typically re…

Knowledge Distillation

Towards Stable Backdoor Purification through Feature Shift Tuning

2023-10-03 · NeurIPS 2023 11 · Rui Min, Zeyu Qin, Li Shen, Minhao Cheng

It has been widely observed that deep neural networks (DNN) are vulnerable to backdoor attacks where attackers could manipulate the model behavior maliciously by tampering with a small set of training samples. Although a…