Gradually Excavating External Knowledge for Implicit Complex Question Answering

TL;DR

This paper introduces a gradual external knowledge excavation framework for open-domain complex question answering, enabling LLMs to iteratively acquire and utilize external information for improved accuracy.

Contribution

The proposed method allows LLMs to actively and iteratively gather external knowledge and reason step-by-step, enhancing performance on complex questions with fewer parameters.

Findings

Achieves 78.17% accuracy on StrategyQA dataset.

Uses less than 6% of parameters compared to competitors.

Sets new state-of-the-art for ~10B-scale LLMs.

Abstract

Recently, large language models (LLMs) have gained much attention for the emergence of human-comparable capabilities and huge potential. However, for open-domain implicit question-answering problems, LLMs may not be the ultimate solution due to the reasons of: 1) uncovered or out-of-date domain knowledge, 2) one-shot generation and hence restricted comprehensiveness. To this end, this work proposes a gradual knowledge excavation framework for open-domain complex question answering, where LLMs iteratively and actively acquire external information, and then reason based on acquired historical knowledge. Specifically, during each step of the solving process, the model selects an action to execute, such as querying external knowledge or performing a single logical reasoning step, to gradually progress toward a final answer. Our method can effectively leverage plug-and-play external knowledge and dynamically adjust the strategy for solving complex questions. Evaluated on the StrategyQA dataset, our method achieves 78.17% accuracy with less than 6% parameters of its competitors, setting new SOTA for ~10B-scale LLMs.