用生成模型修复多模态对抗幻觉，让图像文本对齐更鲁棒
Breaking the Illusion: Consensus-Based Generative Mitigation of Adversarial Illusions in Multi-Modal Embeddings
- 通过变分自编码器生成自然样本，恢复被干扰的跨模态对齐
- 在ImageBind上将攻击成功率降至接近零，提升扰动前后对齐效果
- 无需任务特定设计，适合各类多模态应用防御
多模态基础模型将图像、文本等模态对齐至共享嵌入空间，但仍易受对抗幻觉攻击——微小不可察觉的扰动会破坏跨模态对齐并误导下游任务。为应对该问题，我们提出一种与任务无关的缓解机制：利用生成模型（如变分自编码器，VAEs）净化攻击者扰动的输入，以恢复自然对齐。为进一步增强防御能力，采用生成采样策略结合基于共识的聚合方案，综合多个生成样本结果。在当前领先的多模态编码器ImageBind上的实验表明，该方法显著将幻觉攻击成功率降至接近零，并在未扰动及扰动输入设置下均提升了跨模态对齐性能，提供了一种有效且任务无关的对抗幻觉防御方案。代码已开源：https://github.com/fatemehakb/adversarial-illusions-mitigation。
原文摘要 · Abstract (English)
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 https://github.com/fatemehakb/adversarial-illusions-mitigation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。