A defense that aligns image patches with LLM-generated knowledge elements and downweights poisoned samples reduces backdoor and poisoning attack success in contrastive vision-language models to near zero on tested benchmarks.
CorruptEncoder: Data Poisoning based Backdoor Attacks to Contrastive Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Contrastive learning (CL) pre-trains general-purpose encoders using an unlabeled pre-training dataset, which consists of images or image-text pairs. CL is vulnerable to data poisoning based backdoor attacks (DPBAs), in which an attacker injects poisoned inputs into the pre-training dataset so the encoder is backdoored. However, existing DPBAs achieve limited effectiveness. In this work, we take the first step to analyze the limitations of existing backdoor attacks and propose new DPBAs called CorruptEncoder to CL. CorruptEncoder introduces a new attack strategy to create poisoned inputs and uses a theory-guided method to maximize attack effectiveness. Our experiments show that CorruptEncoder substantially outperforms existing DPBAs. In particular, CorruptEncoder is the first DPBA that achieves more than 90% attack success rates with only a few (3) reference images and a small poisoning ratio 0.5%. Moreover, we also propose a defense, called localized cropping, to defend against DPBAs. Our results show that our defense can reduce the effectiveness of DPBAs, but it sacrifices the utility of the encoder, highlighting the need for new defenses.
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Semantic Shield: Defending Vision-Language Models Against Backdooring and Poisoning via Fine-grained Knowledge Alignment
A defense that aligns image patches with LLM-generated knowledge elements and downweights poisoned samples reduces backdoor and poisoning attack success in contrastive vision-language models to near zero on tested benchmarks.