English(EN) Poisoning Attacks on the PGM-index

PGM-index 学习索引易受导致膨胀的投毒攻击

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-03 04:00

研究人员发现 PGM-index（一种流行的学习索引结构）存在一个重大的漏洞。一种新的攻击（称为 PGM-attack）通过插入对抗性键，可以将索引大小急剧膨胀高达 120 倍。该攻击利用了索引的优化目标，表明即使是理论上最优的结构也可能容易受到投毒攻击。研究结果表明，未来的学习索引需要采用面向鲁棒性的设计。 AI

排序理由该集群包含一篇详细介绍针对特定数据结构的新攻击的研究论文。[lever_c_demoted from research: ic=1 ai=0.7]

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: 16 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍针对特定数据结构的新攻击的研究论文。[lever_c_demoted from research: ic=1 ai=0.7]
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.LG TIER_1 English(EN) · Atsuki Sato, Martin Aum\"uller, Yusuke Matsui · 2026-09-03 04:00

PGM-index上的投毒攻击

arXiv:2609.02328v1 Announce Type: cross Abstract: The PGM-index (Ferragina and Vinciguerra, VLDB'20) is one of the most practical learned indexes, owing to its theoretical elegance and consistently strong empirical performance. It is built on optimal piecewise linear approximatio…

本文如何被排名

报道来源 [1]

PGM-index上的投毒攻击

相关实体

相关话题