Causal Discovery with Language Models as Imperfect Experts

2023/07/05 by Stephanie Long, Alexandre Piché, Long, Stephanie +7 · 21 citations
Computer Science · #Advanced Graph Neural Networks #Artificial Intelligence (cs.AI) #Bayesian Modeling and Causal Inference #Computation and Language (cs.CL) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2307.02390

openalex publication_date 2023/07/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Understanding the causal relationships that underlie a system is a fundamental prerequisite to accurate decision-making. In this work, we explore how expert knowledge can be used to improve the data-driven identification of causal graphs, beyond Markov equivalence classes. In doing so, we consider a setting where we can query an expert about the orientation of causal relationships between variables, but where the expert may provide erroneous information. We propose strategies for amending such expert knowledge based on consistency properties, e.g., acyclicity and conditional independencies in the equivalence class. We then report a case study, on real data, where a large language model is used as an imperfect expert.

Cited by

• CARE: Turning LLMs Into Causal Reasoning Expert
• Bayesian Network Structure Discovery Using Large Language Models
• Linear Causal Discovery with Interventional Constraints
• CaM-Wolf: Causal-Aware Multimodal Agents for Social Deduction Games
• Linear-LLM-SCM: Benchmarking LLMs for Coefficient Elicitation in Linear-Gaussian Causal Models
• Realizing LLMs' Causal Potential Requires Science-Grounded, Novel Benchmarks
• Executable Counterfactuals: Improving LLMs' Causal Reasoning Through Code
• Revealing Multimodal Causality with Large Language Models
• GRID: Graph-based Reasoning for Intervention and Discovery in Built Environments
• CausalMACE: Causality Empowered Multi-Agents in Minecraft Cooperative Tasks
• Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery
• Less Greedy Equivalence Search
• Unveiling Causal Reasoning in Large Language Models: Reality or Mirage?
• Tagged for Direction: Pinning Down Causal Edge Directions with Precision
• Sequential Causal Discovery with Noisy Language Model Priors
• LLM Cannot Discover Causality, and Should Be Restricted to Non-Decisional Support in Causal Discovery
• Causal Inferences from Digital Behavioral Data
• No Black Boxes: Interpretable and Interactable Predictive Healthcare with Knowledge-Enhanced Agentic Causal Discovery
• Ice Cream Doesn't Cause Drowning: Benchmarking LLMs Against Statistical Pitfalls in Causal Inference
• Eroding the Truth-Default: A Causal Analysis of Human Susceptibility to Foundation Model Hallucinations and Disinformation in the Wild
• GoT-CD: Graph-of-Thoughts Causal Discovery and the Fragility of Post-hoc Path-Specific Fairness Audits

Related

• Learning to Defer for Causal Discovery with Imperfect Experts
• Large Language Models for Constrained-Based Causal Discovery
• Causal Order: The Key to Leveraging Imperfect Experts in Causal Inference
• Integrating Large Language Model for Improved Causal Discovery
• Retrieving Classes of Causal Orders with Inconsistent Knowledge Bases