Generative purification with consensus aggregation reduces adversarial illusion attack success rates to near zero on ImageBind while improving alignment on both clean and attacked inputs.
Adversarial-guided diffusion for multimodal llm attacks.arXiv preprint arXiv:2507.23202, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Breaking the Illusion: Consensus-Based Generative Mitigation of Adversarial Illusions in Multi-Modal Embeddings
Generative purification with consensus aggregation reduces adversarial illusion attack success rates to near zero on ImageBind while improving alignment on both clean and attacked inputs.