Skip to content

Translation in progress Some entries still include text in the original language.

Oct 12 · Monday

All papers

Found 12 papers
Filter3

Attack type

Extraction and theft4

Attack technique

Cross-modal payloads3

Impact

Harmful content40Unauthorized actions13Data and privacy leaks7Output manipulation50Availability and reliability5

System type

Models and APIs4LLM applications4AI Agent4Embodied systems and robotics3

Attack surface / entry point

Images, audio and video6

Layer

Application layer6Training and data layer3

System component

MCP and skills21Memory15RAG11Guardrails and judges10Monitors8Reasoning chains10Permissions and sandboxing12Fine-tuning and open weights10Vision12Audio7Video3Image generation5Embeddings5Multimodal26
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 265papers have no tags yet and are not included in filtered results.
Impact：Data and privacy leaksImpact：Availability and reliabilitySystem component：VisionClear all

Another 265papers do not have category tags yet and are not included in filtered results.

  3. OtherarXiv 2610.04792

    多模态一定更好吗？缺失模态鲁棒性的几何视角与 Geodesic Unlearning 方法

    提出 Geodesic Unlearning，沿测地线编辑权重以提升缺模态鲁棒性

    arXiv
    2610.04792
    Published
    Layer
    Model layer
    Setting
    Models and APIs
    Evaluated models
    RoBERTa, ResNet50, Qwen2.5-3B and others6
    AbstractGU 用测地线子空间编辑应对部署时缺模态下降，并尽量保住全模态精度。

    这项工作讨论多模态模型在部署时整段缺一个模态的性能下降，并把校正写成训练后的子空间编辑。

    Paper analysisRead analysis
  4. DefensesarXiv 2610.04616

    PerturBot 用扰动训练打破视觉-语言-动作模型的模态捷径

    提出 Perturbot 扰动训练，打破 VLA 的模态捷径

    arXiv
    2610.04616
    Published
    Layer
    Training and data layer
    Setting
    Embodied systems and robotics
    Evaluated models
    π0.5
    AbstractPerturbot 改训练数据以削弱三类模态捷径，GroundFscore 用来看成功是否来自任务证据。

    Perturbot 是一套不改推理的 VLA 训练期数据干预，用来削弱模态捷径，并用离线分数区分“任务变好”和“靠捷径变好”。

    Paper analysisRead analysis
  8. Evaluation and benchmarksarXiv 2609.33481

    IDUnlearn-Bench：面向视觉语言模型的个体级遗忘基准

    提出 IDUnlearn-Bench，评测 VLM 个体级遗忘后的身份残留

    arXiv
    2609.33481
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    Qwen3-VL-2B, Qwen3-VL-4B, Qwen3-VL-8B and others6
    AbstractIDUnlearn-Bench 用四族探针显示：压低直接属性访问后，身份仍常能被反向查出、绑定或重构。

    IDUnlearn-Bench 评测 VLM 是否忘掉一个可跨模态识别的人，而不只是答不出某条监督事实。

    Paper analysisRead analysis
  10. DefensesarXiv 2609.15671

    QPriv-VL：面向视觉语言模型的问题引导 token 裁剪隐私防御

    提出 QPriv-VL，用问题引导剪枝降低分布式视觉问答的隐私泄露

    arXiv
    2609.15671
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    LLaVA-1.5-7B, Llama-2
    AbstractQPriv-VL 在切层前按问题相关度与 DINOv2 敏感度剪视觉 token，作者称能在约四成 token 下兼顾准确率与四类攻击。

    QPriv-VL 把问题引导的视觉 token 剪枝做成分布式 VQA 的客户端隐私防御，而不是只做集中式推理加速。

    Paper analysisRead analysis
  12. AttacksarXiv 2505.23266

    AdvOF：用对抗物体融合攻击 VLM 驱动的视觉语言导航智能体

    提出 AdvOF，生成对抗物体误导视觉语言导航智能体的场景感知

    arXiv
    2505.23266
    Published
    Layer
    Application layer
    Setting
    Embodied systems and robotics
    Attacker
    White-box
    Evaluated models
    Lseg, Clip, ViT-L/16 and others5
    AbstractAdvOF 以多视角融合生成对抗物体。

    AdvOF 是针对 VLM 驱动 VLN 服务导航的对抗物体攻击。

    Paper analysisRead analysis
You have reached the end