Efficient Causal Graph Discovery Using Large Language Models

TL;DR

This work tackles causal graph discovery with large language models (LLMs) by addressing the quadratic query cost of prior pairwise methods. It introduces a breadth-first search (BFS) prompting framework that constrains the graph to be a DAG and reduces queries to $O(n)$, while optionally incorporating observational statistics through prompts. The approach achieves state-of-the-art or competitive performance on three real-world graphs of varying sizes, including a very large Neuropathic Pain graph where traditional methods fail. The method offers a scalable, data-efficient alternative for causal graph discovery with broad applicability and potential for hybrid integration with statistical methods.

Abstract

We propose a novel framework that leverages LLMs for full causal graph discovery. While previous LLM-based methods have used a pairwise query approach, this requires a quadratic number of queries which quickly becomes impractical for larger causal graphs. In contrast, the proposed framework uses a breadth-first search (BFS) approach which allows it to use only a linear number of queries. We also show that the proposed method can easily incorporate observational data when available, to improve performance. In addition to being more time and data-efficient, the proposed framework achieves state-of-the-art results on real-world causal graphs of varying sizes. The results demonstrate the effectiveness and efficiency of the proposed method in discovering causal relationships, showcasing its potential for broad applicability in causal graph discovery tasks across different domains.

Figures (1)

• Figure 1: Proposed framework for full graph discovery with LLMs