Regularized Multi-LLMs Collaboration for Enhanced Score-based Causal Discovery

TL;DR

This paper introduces a framework that leverages multiple large language models to improve score-based causal discovery from observational data, addressing limitations of prior knowledge and enhancing causal graph reconstruction.

Contribution

It proposes a novel multi-LLM collaboration framework to augment score-based causal discovery methods, utilizing the collective capacity of multiple models.

Findings

Enhanced accuracy in causal graph reconstruction.

Effective utilization of multiple LLMs for causal discovery.

Potential reduction in reliance on prior expert knowledge.

Abstract

As the significance of understanding the cause-and-effect relationships among variables increases in the development of modern systems and algorithms, learning causality from observational data has become a preferred and efficient approach over conducting randomized control trials. However, purely observational data could be insufficient to reconstruct the true causal graph. Consequently, many researchers tried to utilise some form of prior knowledge to improve causal discovery process. In this context, the impressive capabilities of large language models (LLMs) have emerged as a promising alternative to the costly acquisition of prior expert knowledge. In this work, we further explore the potential of using LLMs to enhance causal discovery approaches, particularly focusing on score-based methods, and we propose a general framework to utilise the capacity of not only one but multiple LLMs to augment the discovery process.