Bi-Chainer: Automated Large Language Models Reasoning with Bidirectional Chaining

TL;DR

Bi-Chainer introduces a bidirectional reasoning approach for large language models, improving accuracy and efficiency in solving complex logical problems by dynamically switching reasoning directions.

Contribution

It proposes a novel bidirectional chaining method that enhances logical reasoning in LLMs by leveraging dynamic direction switching and intermediate guidance.

Findings

Achieves significant accuracy improvements over unidirectional methods

Enhances intermediate proof step accuracy

Reduces inference calls for more efficiency

Abstract

Large Language Models (LLMs) have shown human-like reasoning abilities but still face challenges in solving complex logical problems. Existing unidirectional chaining methods, such as forward chaining and backward chaining, suffer from issues like low prediction accuracy and efficiency. To address these, we propose a bidirectional chaining method, Bi-Chainer, which dynamically switches to depth-first reasoning in the opposite reasoning direction when it encounters multiple branching options within the current direction. Thus, the intermediate reasoning results can be utilized as guidance to facilitate the reasoning process. We show that Bi-Chainer achieves sizable accuracy boots over unidirectional chaining frameworks on four challenging logical reasoning datasets. Moreover, Bi-Chainer enhances the accuracy of intermediate proof steps and reduces the average number of inference calls, resulting in more efficient and accurate reasoning.