返回全部动态
LLM增强因果发现：边存在性与方向的概率融合
原标题：LLM-Augmented Causal Discovery: Probabilistic Fusion of Edge Existence and Orientation
AI 摘要
该研究提出概率依赖图（PDG）表示法，将贝叶斯网络结构学习（BNSL）与大型语言模型（LLM）的因果知识进行概率融合。在26个基准网络上，将三种BNSL算法与三种LLM结合，50/50融合在22个网络上优于单一来源，平均F1显著提升。分析显示BNSL提供高召回率的边骨架，而LLM提供高准确率的边方向，两者互补。
以上摘要由 AI 生成，可能存在误差。事实请以原文为准。
正文节选
lightblueRGB198, 226, 255 \jmlrpages LLM-Augmented Causal Discovery: Probabilistic Fusion of Edge Existence and OrientationThanks: Alternative email address: n.k.kitson@qmul.ac.uk Abstract Bayesian network structure learning (BNSL) from observational data struggles with orientation identifiability, while large language models (LLMs) offer broad but often unreliable causal knowledge. We propose combining these complementary sources through a novel representation, termed Probabilistic Dependency G
发布时间：2026-09-01 12:00
抓取时间：2026-08-31 12:11
来源机构：arXiv