PulseAugur
实时 09:02:37
English(EN) Privacy-Preserving Topology-Guided Safety for LLM-Based Multi-Agent Systems via Federated Graph Learning

新的FGLGuard系统通过联邦图学习增强LLM多智能体安全

研究人员开发了FGLGuard,一种用于增强由大型语言模型(LLM)驱动的多智能体系统(MAS)安全性的新颖系统。该方法利用联邦图学习在智能体间通信图上训练图神经网络(GNN),从而能够在不跨组织汇集敏感数据的情况下定位和干预风险智能体。FGLGuard在Agent-SafetyBench、R-Judge和AgentDojo等基准测试中表现出色,优于集中式方法和仅本地训练的方法,同时保持了高可用性和最小的能力损失。 AI

影响 通过实现隐私保护的安全训练,这项研究可能有助于在敏感环境中更安全地部署基于LLM的多智能体系统。

排序理由 该集群包含一篇详细介绍LLM安全新方法的 ist 研究论文。

在 arXiv cs.MA (Multiagent) 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

新的FGLGuard系统通过联邦图学习增强LLM多智能体安全

本文如何被排名

Signal score
1 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群包含一篇详细介绍LLM安全新方法的 ist 研究论文。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, safety, infra
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
2 days old
Coverage has settled into its steady-state source set.

完整方法见我们的编辑标准

报道来源 [2]

  1. arXiv cs.AI TIER_1 English(EN) · Jinxi Yu, Eric Hanchen Jiang, Levina Li, Dong Liu, Zhi Zhang, Wenxiao Zhao, Yanxuan Yu, Kai-Wei Chang, Ying Nian Wu ·

    面向基于LLM的多智能体系统的隐私保护拓扑引导安全:通过联邦图学习实现

    arXiv:2609.02967v1 Announce Type: cross Abstract: Topology-guided safeguards for LLM-based multi-agent systems (MAS) train a GNN over the inter-agent communication graph to localize risky agents and intervene on the topology---but they assume one operator can pool all labeled tra…

  2. arXiv cs.MA (Multiagent) TIER_1 English(EN) · Ying Nian Wu ·

    通过联邦图学习实现基于LLM的多智能体系统的隐私保护拓扑引导安全

    Topology-guided safeguards for LLM-based multi-agent systems (MAS) train a GNN over the inter-agent communication graph to localize risky agents and intervene on the topology---but they assume one operator can pool all labeled traces. Across organizations that assumption breaks: …