English(EN) Published Unlearning Numbers Move Per Checkpoint, and Not Because the Removed Data Survives: An Audit of 263 Released Batch-Normalized Checkpoints

AI 模型遗忘审计发现检查点移动而非数据残留导致数字变化

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-11 04:00

对 263 个已发布的批归一化检查点进行的新审计显示，已发布的遗忘数字可以在检查点之间转移，原因并非被移除的数据得以保留，而是检查点的属性发生了变化。在固定的拟合池中用保留的记录替换被移除的记录，对已发布的单元格影响甚微。然而，检查点已发布状态与重新拟合状态的偏差程度确实与这些转移相关。研究表明，对于批归一化的视觉模型，发布时应在报告的数字旁边注明拟合约定。 AI

影响凸显了验证 AI 模型数据移除的潜在问题，影响了 AI 开发中的信任和合规性。

排序理由该集群包含一篇在 arXiv 上发表的研究论文，详细介绍了对 AI 模型检查点的审计。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 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: 11 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇在 arXiv 上发表的研究论文，详细介绍了对 AI 模型检查点的审计。[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.LG TIER_1 English(EN) · Junlong Shen Xingyu Li · 2026-09-11 04:00

已发布的遗忘数字按检查点移动，并非因为被移除的数据得以保留：对 263 个已发布的批归一化检查点的审计

arXiv:2609.11490v1 Announce Type: cross Abstract: An unlearning audit reads its verdict off numbers that an unlearned model and its retrained reference each publish, and both also ship batch-normalization statistics that no gradient step wrote and no release records. Refitting th…

本文如何被排名

报道来源 [1]

已发布的遗忘数字按检查点移动，并非因为被移除的数据得以保留：对 263 个已发布的批归一化检查点的审计

相关实体

相关话题