aiqbench analysis
This research demonstrates a practical method for enhancing causal graph learning by probabilistically fusing the structural skeletons from traditional Bayesian methods with the orientational insights of Large Language Models. By treating edge states as distributions rather than binary decisions, the approach effectively mitigates the weaknesses of each individual source, achieving statistically significant improvements in F1 scores across diverse benchmarks. The findings highlight a viable pathway for integrating generative AI capabilities into rigorous causal inference tasks without discarding established statistical rigor.
LLM-Augmented Causal Discovery: Probabilistic Fusion of Edge Existence and Orientation
arXiv:2608.27472v1 Announce Type: new 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 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.
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 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.
Artificial Intelligence (cs.AI)
arXiv:2608.27472 [cs.AI]
(or arXiv:2608.27472v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.27472 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 265 words
