Call Me When Necessary: LLMs can Efficiently and Faithfully Reason over Structured Environments

TL;DR

This paper introduces Readi, a novel framework enabling large language models to efficiently and accurately reason over structured environments like knowledge graphs and tables by generating and editing reasoning paths.

Contribution

Readi is the first framework allowing LLMs to generate and iteratively edit reasoning paths for structured environment reasoning, improving efficiency and faithfulness.

Findings

Readi significantly outperforms previous LLM-based methods in KGQA and TableQA tasks.

Readi achieves comparable results to fine-tuned models on some datasets.

Readi substantially enhances vanilla LLM performance on reasoning tasks.

Abstract

Large Language Models (LLMs) have shown potential in reasoning over structured environments, e.g., knowledge graph and table. Such tasks typically require multi-hop reasoning, i.e., match natural language utterance with instances in the environment. Previous methods leverage LLMs to incrementally build a reasoning path, where the LLMs either invoke tools or pick up schemas by step-by-step interacting with the environment. We propose Reasoning-Path-Editing (Readi), a novel framework where LLMs can efficiently and faithfully reason over structured environments. In Readi, LLMs initially generate a reasoning path given a query, and edit the path only when necessary. We instantiate the path on structured environments and provide feedback to edit the path if anything goes wrong. Experimental results on three KGQA and two TableQA datasets show the effectiveness of Readi, significantly surpassing previous LLM-based methods (by 9.1% Hit@1 on WebQSP, 12.4% on MQA-3H and 9.5% on WTQ), comparable with state-of-the-art fine-tuned methods (67% on CWQ and 74.7% on WebQSP) and substantially boosting the vanilla LLMs (by 14.9% on CWQ). Our code will be available on https://aka.ms/readi.