


















Abstract:Multi-modal foundation models align images, text, and other modalities in a shared embedding space but remain vulnerable to adversarial illusions [35], where imperceptible perturbations disrupt cross-modal alignment and mislead downstream tasks. To counteract the effects of adversarial illusions, we propose a task-agnostic mitigation mechanism that purifies the attacker's perturbed input using generative models, e.g., Variational Autoencoders (VAEs), to restore natural alignment. To further enhance the defense mechanism, we adopt a generative sampling strategy combined with a consensus-based aggregation scheme over the outcomes of the generated samples. Our experiments on ImageBind, a state-of-the-art multi-modal encoder, show that our approach substantially reduces the illusion attack success rates to near-zero and improves cross-modal alignment in unperturbed and perturbed input settings, providing an effective and task-agnostic defense against adversarial illusions. The code is available at this https URL.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2511.21893 [cs.LG] |
| (or arXiv:2511.21893v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2511.21893 arXiv-issued DOI via DataCite |
From: Anahita Baninajjar [view email]
[v1]
Wed, 26 Nov 2025 20:18:20 UTC (941 KB)
[v2]
Tue, 21 Apr 2026 12:31:10 UTC (965 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。