DIONYSUS: A Pre-trained Model for Low-Resource Dialogue Summarization

TL;DR

DIONYSUS is a pre-trained encoder-decoder model designed for low-resource dialogue summarization, leveraging pseudo summaries and self-supervised learning to improve performance across domains without extensive labeled data.

Contribution

The paper introduces DIONYSUS, a novel pre-training approach that uses pseudo summaries and dynamic selection to enhance dialogue summarization in low-resource settings.

Findings

DIONYSUS outperforms existing methods on six datasets.

It achieves superior ROUGE scores in zero-shot and few-shot scenarios.

The approach effectively adapts to new domains without labeled data.

Abstract

Dialogue summarization has recently garnered significant attention due to its wide range of applications. However, existing methods for summarizing dialogues have limitations because they do not take into account the inherent structure of dialogue and rely heavily on labeled data, which can lead to poor performance in new domains. In this work, we propose DIONYSUS (dynamic input optimization in pre-training for dialogue summarization), a pre-trained encoder-decoder model for summarizing dialogues in any new domain. To pre-train DIONYSUS, we create two pseudo summaries for each dialogue example: one is produced by a fine-tuned summarization model, and the other is a collection of dialogue turns that convey important information. We then choose one of these pseudo summaries based on the difference in information distribution across different types of dialogues. This selected pseudo summary serves as the objective for pre-training DIONYSUS using a self-supervised approach on a large dialogue corpus. Our experiments show that DIONYSUS outperforms existing methods on six datasets, as demonstrated by its ROUGE scores in zero-shot and few-shot settings.