English(EN) CUSE: Fine-Grained Multimodal Unlearning with Counterfactual Guidance and On-Policy Self-Distillation

新的CUSE框架实现了多模态大语言模型（MLLM）的细粒度遗忘

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-15 04:00

研究人员开发了CUSE，一种用于多模态大语言模型（MLLM）细粒度遗忘的新颖框架。该方法侧重于选择性地删除敏感或受版权保护的信息，同时保留其他内容，即使是在同一响应中。CUSE利用反事实输入来精确识别需要更强遗忘的特定标记，并采用同策略自蒸馏将遗忘和保留目标与模型的当前行为对齐。在MLLMU-Bench上的评估证明了CUSE的有效性，显著减少了目标信息的遗忘，同时提高了非目标内容的保留率。 AI

影响引入了一种更精确的方法来控制大语言模型（LLM）中的信息删除，可能提高数据隐私和安全性。

排序理由关于LLM遗忘新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 3 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 关于LLM遗忘新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, safety; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Chenlu Ding, Jiancan Wu, Yanchen Luo, Tianyu Zhang, Yancheng Yuan, Xiang Wang, Xiangnan He · 2026-10-15 04:00

CUSE：具有反事实指导和同策略自蒸馏的细粒度多模态遗忘

arXiv:2610.14763v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) can reproduce sensitive, copyrighted, or unsafe information from their training data. Effective unlearning must remove target information while preserving non-target content, even within the …

本文如何被排名

报道来源 [1]

CUSE：具有反事实指导和同策略自蒸馏的细粒度多模态遗忘

相关实体

相关话题