Few-shot Query-Focused Summarization with Prefix-Merging

TL;DR

This paper introduces prefix-merging, a prefix-based pretraining strategy that leverages knowledge from text summarization and question answering to improve few-shot query-focused summarization, outperforming traditional fine-tuning methods.

Contribution

The paper proposes a novel prefix-merging pretraining approach that effectively transfers knowledge from related tasks to enhance few-shot query-focused summarization.

Findings

Prefix-merging outperforms fine-tuning in few-shot settings.

Different prefix designs influence the effectiveness of the method.

A visualized explanation clarifies how prefix-merging works.

Abstract

Query-focused summarization has been considered as an important extension for text summarization. It aims to generate a concise highlight for a given query. Different from text summarization, query-focused summarization has long been plagued by the problem of lacking high-quality large-scale datasets. In this paper, we investigate the idea that whether we can integrate and transfer the knowledge of text summarization and question answering to assist the few-shot learning in query-focused summarization. Here, we propose prefix-merging, a prefix-based pretraining strategy for few-shot learning in query-focused summarization. Drawn inspiration from prefix-tuning, we are allowed to integrate the task knowledge from text summarization and question answering into a properly designed prefix and apply the merged prefix to query-focused summarization. With only a small amount of trainable parameters, prefix-merging outperforms fine-tuning on query-focused summarization. We further discuss the influence of different prefix designs and propose a visualized explanation for how prefix-merging works.