Skip to content

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

Oct 12 · Monday

All papers

Found 9 papers
Filter4

Attack type

Poisoning and backdoors3

Attack technique

Gradient and representation optimization3

Defense type

Model hardening3

Impact

Harmful content76Unauthorized actions12Data and privacy leaks23Output manipulation66Availability and reliability7Oversight failure6Increased misuse capability2

System type

Models and APIs6AI Agent3Embodied systems and robotics3

Attack surface / entry point

User input11Web pages and interfaces3Tool and skill descriptions3Images, audio and video5Training data8Weights and gradients5Distribution channels4

Layer

Training and data layer5

System component

Fine-tuning and open weights3Multimodal4
Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 233papers have no tags yet and are not included in filtered results.
Impact：Availability and reliabilityImpact：Increased misuse capabilityAttack surface / entry point：Images, audio and videoAttack surface / entry point：Weights and gradientsClear all

Another 233papers do not have category tags yet and are not included in filtered results.

  1. AttacksarXiv 2610.09819

    FAB 攻击：给 HuBERT/WavLM 植入后门，背景警笛声即可让下游语音任务失灵

    提出 FAB 攻击，向声学基础模型植入可由环境声触发的后门

    arXiv
    2610.09819
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    HuBERT-base, WavLM-base

    Abstract论文提出了针对声学基模的 FAB 后门攻击，在弱威胁模型下实现了任务无关、物理可触发的后门植入。

    Paper analysisRead analysis
  2. Analysis and theoryarXiv 2610.09004

    研究：去除有害信息不足以证明开源权重模型的抗篡改能力

    论证发布时互信息不能认证开源权重抗微调恢复

    arXiv
    2610.09004
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Evaluated models
    EleutherAI/deep-ignorance-e2e-strong-filter, EleutherAI/deep-ignorance-unfiltered, Google BERT-Tiny and others6
    Abstract作者称零信息仍可一步恢复，表观抗微调可来自参数化。

    作者研究开源权重模型的篡改抵抗：发布时的互信息是否足以保证微调恢复很慢。

    Paper analysisRead analysis
  4. AttacksarXiv 2610.06083

    研究者提出面向深度强化学习的轨迹级可迁移黑盒对抗攻击

    提出轨迹级可迁移对抗攻击，降低深度强化学习策略的期望回报

    arXiv
    2610.06083
    Published
    Layer
    Model layer
    Setting
    Embodied systems and robotics
    Evaluated models
    DQN, Double DQN (DDQN)
    Abstract轨迹级后退视界攻击在 CartPole 上比逐步 FGSM 类和 Sign 噪声更压低回报。

    这项工作研究 DRL 上的迁移式黑盒观测攻击：受害参数不可访问时，在白盒代理上造扰动再迁移。

    Paper analysisRead analysis
  5. DefensesarXiv 2609.39866

    研究者提出梯度封堵实现 LLM 发布前的预防性遗忘

    提出 GSU 在发布前封堵敏感门梯度，抵抗下游微调获取禁止能力

    arXiv
    2609.39866
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Attacker
    White-box
    Evaluated models
    Llama-3.2-1B, Llama-3.2-3B, Llama-3.1-8B and others7
    AbstractGSU 在发布前封堵 SiLU 门的获取梯度，并在 TOFU 与 WMDP 上降低下游微调得分。

    Gradient-Sealed Unlearning（GSU）是一种发布前的 preemptive unlearning：不删除模型里已经有的能力，而是让指定禁止域能力更难被发布后的微调获取。

    Paper analysisRead analysis
  6. AttacksarXiv 2609.27124

    Fed-ADR：用恶意编排服务器协同客户端梯度操纵攻击联邦学习

    提出 Fed-ADR，用编排服务器协同操纵梯度攻破联邦学习并检测恢复

    arXiv
    2609.27124
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    Attacker
    Gray-box
    Evaluated models
    CNN（MNIST：2 Conv + 3 FC）, CNN（Fashion-MNIST：6 Conv + 1 FC）, ResNet18（CIFAR-10）
    AbstractFed-ADR 用 gray-box 编排攻击绕过多种聚合防御，再用 Hessian 一致性检测和窗口内恢复把准确率拉回。

    Fed-ADR 是一套面向联邦学习的攻击、检测与恢复框架：编排服务器改写合谋客户端的梯度，使参数服务器的现有防御把投毒更新当成良性，同时在发现后用历史梯度把全局模型原地修回来。

    Paper analysisRead analysis
  7. AttacksarXiv 2609.01495

    去中心化联邦学习中的拜占庭节点放置优化

    提出 BPI 选点算法，增强去中心化联邦学习中的模型与数据投毒

    arXiv
    2609.01495
    Published
    Layer
    Training and data layer
    Setting
    Models and APIs
    AbstractBPI 用有限轮 gossip 暴露选拜占庭集合。

    作者把 DFL 中“破坏哪 m 个节点”写成攻击者在固定预算下的优化问题，并用 BPI 近似有限轮内诚实节点受到的拜占庭暴露。

    Paper analysisRead analysis
  8. Evaluation and benchmarksarXiv 2602.10639

    VideoSTF 揭示视频大模型输出重复问题

    提出 VIDEO STF 评测视频大模型输出重复并压力测试时序扰动

    arXiv
    2602.10639
    Published
    Layer
    Prompt layer
    Setting
    Models and APIs
    Evaluated models
    LLaVA-Video-7B-Qwen2, LLaVA-Video-7B-Qwen2-Video-Only, LLaVA-NeXT-Video-7B and others10
    AbstractVIDEO STF 把输出重复做成可测、可被时间扰动放大、并可黑盒诱导的失败。

    VIDEO STF 测量 VideoLLM 的输出重复，并把时间变换用作 black-box 诱导。

    Paper analysisRead analysis
  9. 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