Log In Sign Up

Strong Baseline Defenses Against Clean-Label Poisoning Attacks

by   Neal Gupta, et al.

Targeted clean-label poisoning is a type of adversarial attack on machine learning systems where the adversary injects a few correctly-labeled, minimally-perturbed samples into the training data thus causing the deployed model to misclassify a particular test sample during inference. Although defenses have been proposed for general poisoning attacks (those which aim to reduce overall test accuracy), no reliable defense for clean-label attacks has been demonstrated, despite the attacks' effectiveness and their realistic use cases. In this work, we propose a set of simple, yet highly-effective defenses against these attacks. We test our proposed approach against two recently published clean-label poisoning attacks, both of which use the CIFAR-10 dataset. After reproducing their experiments, we demonstrate that our defenses are able to detect over 99 them without any compromise on model performance. Our simple defenses show that current clean-label poisoning attack strategies can be annulled, and serve as strong but simple-to-implement baseline defense for which to test future clean-label poisoning attacks.


page 1

page 2

page 3

page 4


Narcissus: A Practical Clean-Label Backdoor Attack with Limited Information

Backdoor attacks insert malicious data into a training set so that, duri...

Traceback of Data Poisoning Attacks in Neural Networks

In adversarial machine learning, new defenses against attacks on deep le...

Turning a Curse Into a Blessing: Enabling Clean-Data-Free Defenses by Model Inversion

It is becoming increasingly common to utilize pre-trained models provide...

Deep Partition Aggregation: Provable Defense against General Poisoning Attacks

Adversarial poisoning attacks distort training data in order to corrupt ...

Triggerless Backdoor Attack for NLP Tasks with Clean Labels

Backdoor attacks pose a new threat to NLP models. A standard strategy to...

Low-Loss Subspace Compression for Clean Gains against Multi-Agent Backdoor Attacks

Recent exploration of the multi-agent backdoor attack demonstrated the b...

Classification Auto-Encoder based Detector against Diverse Data Poisoning Attacks

Poisoning attacks are a category of adversarial machine learning threats...