大语言模型增强的因果发现：边存在性与方向的概率融合
LLM-Augmented Causal Discovery: Probabilistic Fusion of Edge Existence and Orientation
浏览论文内容
中文总结 AI 辅助
该研究将BNSL与Gemini等LLM的因果知识通过概率依赖图融合，在26个基准网络上使F1值平均提升0.056，证明概率不确定性表示可有效提升因果图准确性。
中文摘要 AI 辅助
从观测数据中进行贝叶斯网络结构学习（BNSL）面临方向可识别性的难题，而大语言模型（LLM）能提供广泛但通常不可靠的因果知识。我们提出通过一种名为概率依赖图（PDG）的新型表示来结合这两种互补来源。在PDG中，每条边与有向、无向及缺失状态的分布相关联，可通过加权平均实现融合。我们在26个基准网络上评估该方法，将三种BNSL算法（FGES、Tabu、PC）的集成与三种LLM（Gemini、Claude、GPT）在多个提示词及随机种子下结合。简单的50/50融合在26个网络中的22个上，其F1值优于单独使用任一来源，具有统计学意义的平均提升为0.056（p<0.001）。分析显示两种来源发挥互补作用：BNSL贡献高召回率的边骨架（80%，而LLM为60%），LLM贡献准确的边方向（96%，而BNSL为77%）。我们的结果表明，将两种来源表示为边存在性与方向上的概率不确定性，是提升因果图准确性的实用且有效方法。
英文摘要
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 Graphs (PDGs). In a PDG, each edge is associated with a distribution over directed, undirected, and absent states, enabling fusion via weighted averaging. We evaluate this approach on 26 benchmark networks, combining ensembles of three BNSL algorithms (FGES, Tabu, PC) with three LLMs (Gemini, Claude, GPT) across multiple prompts and random seeds. A simple 50/50 fusion improves F1 over the better of either source alone in 22 of 26 networks, with a statistically significant mean improvement of $0.056$ $(p<0.001)$. Analysis reveals that the two sources play complementary roles: BNSL contributes a high-recall edge skeleton (80\% vs 60\% for LLM), while LLM contributes accurate edge orientation (96\% vs 77\% for BNSL). Our results show that representing both sources as probabilistic uncertainty over edge existence and orientation is a practical and effective way to improve causal graph accuracy.
发表机构
- Queen Mary University of London（伦敦玛丽女王大学）
- Causaliq
机构由 AI 辅助整理，请以论文原文为准。
