English(EN) Auditable Unlearning for LLMs

提出 LLM 新的“可审计遗忘”指标

作者 PulseAugur 编辑部 · [1 个来源] · 2026-11-05 05:00

研究人员引入了“可审计遗忘”作为评估机器学习遗忘技术的新指标，重点关注模型提供者提供数据删除证据的能力。当前方法在四个关键维度上存在不足：遗忘性、能力保留性、鲁棒性和可审计性。为解决此问题，提出了一种名为 GapWiden 和 GapVeil 的两阶段方法，旨在增加已删除数据和保留数据审计统计数据之间的差异，并使用拒绝采样器来确保可审计性。该方法表明，删除证据可以被设计到遗忘方法中，可能改变现有技术的可行性。 AI

影响为 LLM 遗忘引入了新的评估维度，可能影响 AI 系统的数据隐私和合规性。

排序理由学术论文，介绍新概念和方法。[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: 26 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 学术论文，介绍新概念和方法。[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: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Piyush Tiwary, Anushka Srivastava, Prathosh A. P · 2026-11-05 05:00

LLM 的可审计遗忘

arXiv:2611.05866v1 Announce Type: new Abstract: Machine unlearning aims to remove a user's data from a trained model after a deletion request. Existing methods are majorly evaluated on three axes - how well they forget, how much capability they retain, and how robust they are to …

本文如何被排名

报道来源 [1]

LLM 的可审计遗忘

相关实体

相关话题