PulseAugur
中
实时 11:08:41
English(EN) Ctrl-G: Controlling Structural and Interaction Exposure in Knowledge Graph Augmented LLM Reasoning

新的Ctrl-G框架通过知识图谱增强LLM隐私

作者 PulseAugur 编辑部 · [1 个来源] ·

研究人员开发了Ctrl-G，一个新颖的框架，旨在增强知识图谱（KGs）与大型语言模型（LLMs）结合使用时的隐私性。该KG-RAG系统通过在将图谱发送到远程LLM之前，在本地创建条件化查询的、经过清理的图谱视图来解决结构和交互暴露问题。Ctrl-G采用分层控制器进行规划和探索，利用经验记忆来减少跨多个查询的重复披露。该框架表现强劲，在六个基准测试中的五个上达到了最高准确率，并且比最强的基线高出13.8个百分点。 AI

影响 增强了LLM在知识图谱上推理的隐私性，可能支持更安全地使用敏感数据。

排序理由 该集群包含一篇详细介绍知识图谱增强LLM推理新框架的研究论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

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

新的Ctrl-G框架通过知识图谱增强LLM隐私

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
Tool
该集群包含一篇详细介绍知识图谱增强LLM推理新框架的研究论文。[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
1 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 [1]

  1. arXiv cs.CL TIER_1 English(EN) · Xingyu Tan, Xiaoyang Wang, Qing Liu, Xiwei Xu, Xin Yuan, Liming Zhu, Wenjie Zhang ·

    Ctrl-G: Controlling Structural and Interaction Exposure in Knowledge Graph Augmented LLM Reasoning

    arXiv:2601.08739v2 Announce Type: replace Abstract: Knowledge graphs (KGs) provide structured evidence for Web question answering with large language models (LLMs), but sending retrieved evidence to a remote LLM can expose sensitive graph information. Existing privacy treatments …

相关实体

相关话题