Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 240papers have no tags yet and are not included in filtered results.

System type：Models and APIs Layer：Application layer Layer：Training and data layer System component：Embeddings System component：Multimodal Clear all

Another 240papers do not have category tags yet and are not included in filtered results.

Evaluation and benchmarksarXiv 2610.10358Oct 8
Open-MMUnlearning 发布：统一多模态大模型遗忘方法与评测框架
提出统一框架评测多模态大模型的遗忘方法
arXiv
2610.10358
Published
Oct 8
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
LLaVA-1.5-7B
- DefensesModel hardening
- ComponentVision
Paper analysisRead analysis
DefensesarXiv 2610.09941Oct 7
OrthoPurify 用一步正交投影净化被后门污染的视觉语言模型
提出 OrthoPurify，用正交投影清除视觉语言模型适配器中的劫持方向
arXiv
2610.09941
Published
Oct 7
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
LLaVA-1.5-7B, Qwen3-VL-8B, LLaVA-1.5-13B and others5
- DefensesModel hardening
- AttacksPoisoning and backdoors
- ComponentFine-tuning and open weights
Abstract一次投影去掉劫持方向。
OrthoPurify 是部署前的 LVLM 后门净化方法，用一次正交投影去掉被劫持的权重更新方向。
Paper analysisRead analysis
AttacksarXiv 2610.09819Oct 7
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
Oct 7
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
HuBERT-base, WavLM-base
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentAudio
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
DefensesarXiv 2610.05637Oct 5
Study shows VLM visual grounding refusal rates are on average 31 to 59 percentage points lower by harm domain
提出融合定位拒答与能力保留的安全微调，补齐视觉定位输出的拒答缺口
arXiv
2610.05637
Published
Oct 5
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Claude Sonnet 4.6, Gemini 3.5 Flash, Molmo2-8B and others5
- DefensesModel hardening
- RiskInappropriate refusal
- ComponentVision
Abstract论文揭示了 VLM 定位安全落后于 VQA 的对齐断层，提出三成分微调方案有效弥合差距并保留定位能力，表征分析揭示了对齐机制。
Paper analysisRead analysis
AttacksarXiv 2610.05453Oct 5
研究者提出 Privacy-Leaking Watermarks，可在本地多模态模型中泄露对话隐私
提出 PLW，用微调把先前对话敏感触发编码进后续生成图像
arXiv
2610.05453
Published
Oct 5
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
OmniGen2, BAGEL-7B
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentFine-tuning and open weights
AbstractPLW 把先前对话线索写成后续图像中的隐水印，检出随触发、模型和间隔变化。
Paper analysisRead analysis
DefensesarXiv 2610.05401Oct 5
AVCE：面向扩散模型可验证概念擦除的审计感知遗忘框架
提出审计感知遗忘框架 AVCE，从文本与视觉通路擦除扩散模型概念
arXiv
2610.05401
Published
Oct 5
Layer
Training and data layer
Setting
Models and APIs
Attacker
White-box, Black-box
Evaluated models
Stable Diffusion XL (SDXL), Stable Diffusion v1.5 (SD v1.5), Flux 1.0
- DefensesModel hardening
- AttacksJailbreaks
- ComponentImage generation
AbstractAVCE 在审计锚点上改两条注意力通路。
AVCE 是面向扩散模型的概念擦除方法，要让擦除经得起绕过文本条件的潜空间审计，而不只是挡住对抗提示词。
Paper analysisRead analysis
OtherarXiv 2610.05264Oct 4
面向音频深度伪造检测的任务感知联合剪枝与蒸馏框架
提出任务感知的跨域蒸馏与剪枝方法以压缩音频深度伪造检测器
arXiv
2610.05264
Published
Oct 4
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
XLSR-AASIST
- ComponentAudio
Paper analysisRead analysis
DefensesarXiv 2610.03980Oct 3
FADE：面向视频扩散模型的帧感知多概念擦除框架
提出 FADE，对文生视频模型做帧感知多概念擦除
arXiv
2610.03980
Published
Oct 3
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Wan2.1-T2V-1.3B, Wan2.1-T2V-14B, CogVideoX-2B and others4
- DefensesModel hardening
- AttacksMalicious use
- ComponentImage generation
AbstractFADE 用闭式编辑、帧感知适配器和软路由，在单骨干上擦除多概念并压低逐帧残差。
FADE 是面向文生视频扩散 Transformer 的多概念遗忘方法，用来抑制指定物体、画风、裸体、暴力、仇恨和名人身份，并尽量保持其他生成。作者要处理两件事。与帧无关的抑制会让概念在少数帧重现，片段均值会掩盖这一失败。
Paper analysisRead analysis
DefensesarXiv 2610.02967Oct 2
用偏好奖励与评分标准奖励组合后训练前沿文生图模型
组合偏好奖励与评分标准奖励后训练文生图模型以抑制奖励作弊
arXiv
2610.02967
Published
Oct 2
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
FLUX.2-dev, Ideogram-4
- DefensesModel hardening
- RiskReward hacking
- ComponentImage generation
Abstract组合奖励后训练 FLUX.2-dev 与 Ideogram-4，并报告高于基座的胜率与 Elo。
作者给出一套开放域文生图后训练配方：把人类偏好奖励和 prompt 条件的 rubric 奖励合在一起，而不是换一个新优化器。问题是单一奖励盖不住用户意图。偏好模型能看整体观感，但对“是否实现每个要求”这类稀疏属性不敏感。
Paper analysisRead analysis
DefensesarXiv 2610.01800Oct 1
LineupRL：基于描述与时序配对识别的可验证强化学习时间序列描述方法
提出 LineupRL，用描述到序列识别作可验证奖励训练时序描述模型
arXiv
2610.01800
Published
Oct 1
Layer
Training and data layer
Setting
Models and APIs
- DefensesModel hardening
- RiskReward hacking
- ComponentVision
AbstractLineupRL 用统计近邻上的序列识别做可验证奖励，3B 描述器超过 SFT、两种 RL 奖励和 72B 教师。
LineupRL 是给开放式时间序列描述用的 RLVR：不模仿合成参考，而奖励一条描述能否让冻结文本 LLM 从相近序列里认出目标。
Paper analysisRead analysis
OtherarXiv 2610.00812Oct 1
视频生成模型的后训练与对齐综述
综述视频生成的后训练与对齐，并汇总基准与开放挑战
arXiv
2610.00812
Published
Oct 1
Layer
Training and data layer
Setting
Models and APIs
- ComponentImage generation
Abstract综述把视频后训练分成四类对齐信号，并指出奖励、长视频、权衡和安全仍开放。
这是一篇视频生成 后训练与对齐 综述：预训练模型已有强先验，但作者认为仍不能可靠遵循意图、维持时间连贯或满足物理与安全约束。
Paper analysisRead analysis
AttacksarXiv 2609.39047Sep 30
BadAction：通过动作引导触发器对交互式视频生成模型实施后门攻击
提出 BadAction，通过动作触发器给交互式视频生成植入后门
arXiv
2609.39047
Published
Sep 30
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Astra (Wan2.1-T2V-1.3B)
- AttacksPoisoning and backdoors
- TechniqueDormant triggers
- ComponentFine-tuning and open weights
Paper analysisRead analysis
DefensesarXiv 2609.37837Sep 29
Intent-Privilege OPSD：让视觉语言模型在隐式风险下兼顾安全与有用
提出 Intent-Privilege OPSD，对齐视觉语言模型在隐式风险下的安全与帮助性
arXiv
2609.37837
Published
Sep 29
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Qwen3-VL-4B-Instruct, LLaVA, LLaVA-1.5-7B and others4
- DefensesModel hardening
- RiskInappropriate refusal
- ComponentVision
Abstract作者报告训练期证据化意图蒸馏兼顾安全与帮助性，效果受标注质量和评测设置影响。
Intent-Privilege OPSD 是针对跨模态隐式风险的训练期安全—帮助性对齐方法。作者关注的不是直接拒绝敏感请求，而是按可观察图文证据保留正常帮助、约束有害信息。
Paper analysisRead analysis
DefensesarXiv 2609.36099Sep 29
通过逆向蒸馏实现数据遗忘（IDU）
提出 IDU，在无保留样本时蒸馏单步生成器并抑制遗忘子集
arXiv
2609.36099
Published
Sep 29
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
IDU one-step generator
- DefensesModel hardening
- ComponentImage generation
AbstractIDU 同时做单步蒸馏和数据遗忘，在 MNIST 与 CIFAR-10 上压低遗忘类生成率。
IDU 面向无条件 flow 与 score-based 教师，在蒸馏成单步生成器的同时做数据遗忘。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.34092Sep 28
机器遗忘在 ASR 中是否有效：arXiv 论文评估现有算法与评测工具
arXiv
2609.34092
Published
Sep 28
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
E-Branchformer
- DefensesModel hardening
- AttacksExtraction and theft
- ComponentAudio
Abstract梯度上升类方法更接近简单评测下的重训参照，复杂方法会被知情攻击认出，多说话人更差。
Paper analysisRead analysis
Evaluation and benchmarksarXiv 2609.33481Sep 27
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
Sep 27
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Qwen3-VL-2B, Qwen3-VL-4B, Qwen3-VL-8B and others6
- DefensesModel hardening
- AttacksExtraction and theft
- ComponentVision
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
AttacksarXiv 2609.33445Sep 27
CSR：跨架构统一攻击概念擦除
提出跨架构攻击 CSR，用概念分数参数重学习恢复文生图被擦除概念
arXiv
2609.33445
Published
Sep 27
Layer
Training and data layer
Setting
Models and APIs
Attacker
White-box
Evaluated models
FLUX.1-dev, Stable Diffusion-v1.4
- AttacksModel tampering
- ComponentImage generation
Paper analysisRead analysis
DefensesarXiv 2609.31911Sep 26
用强化学习增强胸部 X 光报告生成中的视觉推理
用门控 IoU 与 RadGraph 的 GRPO 降低胸片报告幻觉
arXiv
2609.31911
Published
Sep 26
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Qwen3-VL-8B-GRPO
- DefensesModel hardening
- RiskHallucination and fabrication
- ComponentVision
Paper analysisRead analysis
OtherarXiv 2609.29173Sep 24
透过人眼与机器眼：理解视频透视扩展现实中的视图错配
形式化视频透视头显中系统截图与人眼视野的错配
arXiv
2609.29173
Published
Sep 24
Layer
Application layer
Setting
Models and APIs
Evaluated models
GPT-5.4
- ComponentVision
Paper analysisRead analysis
InterpretabilityarXiv 2609.24243Sep 21
TAME：用 SAE 抑制激活缓解 VLM 思维链混淆
提出 TAME，在强化学习中抑制模板激活以缓解思维链混淆
arXiv
2609.24243
Published
Sep 21
Layer
Training and data layer
Setting
Models and APIs
Evaluated models
Qwen3-VL-8B, Gemma-3-12B-it
- DefensesModel hardening
- ComponentReasoning chains
Paper analysisRead analysis

AI Safety Daily

Read the latest issueOne issue per day, updated as events develop