Step-by-Step Causality: Transparent Causal Discovery with Multi-Agent Tree-Query and Adversarial Confidence Estimation

TL;DR

This paper presents Tree-Query, a multi-agent LLM framework for transparent causal discovery that provides interpretable, confidence-rated causal judgments without data, improving over baseline methods and enabling causal priors extraction.

Contribution

Introduces Tree-Query, a novel multi-agent LLM approach that reduces causal discovery to interpretable queries with confidence scores, backed by theoretical guarantees.

Findings

Improves structural metrics over baseline LLM methods on benchmark datasets

Provides asymptotic guarantees for causal relation identifiability

Demonstrates practical utility with a case study on confounder screening

Abstract

Causal discovery aims to recover ``what causes what'', but classical constraint-based methods (e.g., PC, FCI) suffer from error propagation, and recent LLM-based causal oracles often behave as opaque, confidence-free black boxes. This paper introduces Tree-Query, a tree-structured, multi-expert LLM framework that reduces pairwise causal discovery to a short sequence of queries about backdoor paths, (in)dependence, latent confounding, and causal direction, yielding interpretable judgments with robustness-aware confidence scores. Theoretical guarantees are provided for asymptotic identifiability of four pairwise relations. On data-free benchmarks derived from Mooij et al. and UCI causal graphs, Tree-Query improves structural metrics over direct LLM baselines, and a diet--weight case study illustrates confounder screening and stable, high-confidence causal conclusions. Tree-Query thus offers a principled way to obtain data-free causal priors from LLMs that can complement downstream data-driven causal discovery. Code is available at https://anonymous.4open.science/r/Repo-9B3E-4F96.