Leveraging large language models for causal discovery: a constraint-based, argumentation-driven approach
File(s)
Accepted version
Author(s)
Li, Zihao
Russo, Fabrizio
Type
Conference Paper
Abstract
Causal discovery seeks to uncover causal relations from data, typically represented as causal graphs, and is essential for predicting the effects of interventions. While expert knowledge is required to construct principled causal graphs, many statistical methods have been proposed to leverage observational data with varying formal guarantees. Causal Assumption-based Argumentation (ABA) is a framework that uses symbolic reasoning to ensure correspondence between input constraints and output graphs, while offering a principled way to combine data and expertise. We explore the use of large language models (LLMs) as imperfect experts, eliciting semantic structural constraints from variable names and descriptions and integrating them with statistical evidence through Causal ABA. We propose ABAPC-LLM as a principled and conservative approach to hybrid data- and LLM-driven causal discovery. Additionally, we introduce an evaluation protocol to mitigate memorisation bias when assessing LLMs for causal discovery and show competitive performance on novel semantically grounded random benchmarks, as well as on standard small- and medium-sized benchmarks.
Date Issued
2026-08-01
Date Acceptance
2026-06-01
Citation
Proceedings of Machine Learning Research, 2026, pp.3631-3667
ISSN
2640-3498
Publisher
MLResearchPress
Start Page
3631
End Page
3667
Journal / Book Title
Proceedings of Machine Learning Research
Copyright Statement
Copyright © 2026 Association for Uncertainty in Artificial Intelligence. This is the author’s accepted manuscript made available under a CC-BY licence in accordance with Imperial’s Research Publications Open Access policy (www.imperial.ac.uk/oa-policy)
License URL
Source
42nd Conference on Uncertainty in Artificial Intelligence (uai 2026)
Publication Status
Published
Start Date
2026-08-17
Finish Date
2026-08-21
Coverage Spatial
Amsterdam, Netherlands
Date Publish Online
2026-09-02