Multi-Document Summarization with Centroid-Based Pretraining

TL;DR

This paper introduces a novel pretraining method for multi-document summarization that uses ROUGE-based centroids as proxies for summaries, enabling effective training without human summaries and achieving competitive results.

Contribution

The paper proposes a centroid-based pretraining objective for MDS that does not require human summaries, improving performance across various datasets.

Findings

Centrum outperforms or matches state-of-the-art models.

Pretraining on document sets alone is effective.

Models are publicly available for research use.

Abstract

In Multi-Document Summarization (MDS), the input can be modeled as a set of documents, and the output is its summary. In this paper, we focus on pretraining objectives for MDS. Specifically, we introduce a novel pretraining objective, which involves selecting the ROUGE-based centroid of each document cluster as a proxy for its summary. Our objective thus does not require human written summaries and can be utilized for pretraining on a dataset consisting solely of document sets. Through zero-shot, few-shot, and fully supervised experiments on multiple MDS datasets, we show that our model Centrum is better or comparable to a state-of-the-art model. We make the pretrained and fine-tuned models freely available to the research community https://github.com/ratishsp/centrum.