Prompt-based Conservation Learning for Multi-hop Question Answering

TL;DR

This paper introduces a prompt-based conservation learning framework for multi-hop question answering that effectively acquires new reasoning skills while retaining knowledge from single-hop tasks, improving performance and interpretability.

Contribution

It proposes a novel framework that combines prompt-based learning with conservation strategies to enhance multi-hop QA without forgetting prior knowledge.

Findings

PCL achieves competitive results on HotpotQA benchmark.

PCL retains high performance on single-hop sub-questions.

PCL effectively mitigates forgetting in multi-task QA settings.

Abstract

Multi-hop question answering (QA) requires reasoning over multiple documents to answer a complex question and provide interpretable supporting evidence. However, providing supporting evidence is not enough to demonstrate that a model has performed the desired reasoning to reach the correct answer. Most existing multi-hop QA methods fail to answer a large fraction of sub-questions, even if their parent questions are answered correctly. In this paper, we propose the Prompt-based Conservation Learning (PCL) framework for multi-hop QA, which acquires new knowledge from multi-hop QA tasks while conserving old knowledge learned on single-hop QA tasks, mitigating forgetting. Specifically, we first train a model on existing single-hop QA tasks, and then freeze this model and expand it by allocating additional sub-networks for the multi-hop QA task. Moreover, to condition pre-trained language models to stimulate the kind of reasoning required for specific multi-hop questions, we learn soft prompts for the novel sub-networks to perform type-specific reasoning. Experimental results on the HotpotQA benchmark show that PCL is competitive for multi-hop QA and retains good performance on the corresponding single-hop sub-questions, demonstrating the efficacy of PCL in mitigating knowledge loss by forgetting.