TWAG: A Topic-Guided Wikipedia Abstract Generator

TL;DR

This paper introduces TWAG, a two-stage topic-guided model for Wikipedia abstract generation that improves the quality of summaries by incorporating topical information and decomposing content into different topics.

Contribution

The paper proposes a novel two-stage model that detects topics in input paragraphs and uses topic-aware representations to generate more comprehensive Wikipedia abstracts.

Findings

TWAG outperforms existing baselines on the WikiCatSum dataset.

The model effectively decomposes abstracts into different topics.

TWAG generates more comprehensive and accurate summaries.

Abstract

Wikipedia abstract generation aims to distill a Wikipedia abstract from web sources and has met significant success by adopting multi-document summarization techniques. However, previous works generally view the abstract as plain text, ignoring the fact that it is a description of a certain entity and can be decomposed into different topics. In this paper, we propose a two-stage model TWAG that guides the abstract generation with topical information. First, we detect the topic of each input paragraph with a classifier trained on existing Wikipedia articles to divide input documents into different topics. Then, we predict the topic distribution of each abstract sentence, and decode the sentence from topic-aware representations with a Pointer-Generator network. We evaluate our model on the WikiCatSum dataset, and the results show that \modelnames outperforms various existing baselines and is capable of generating comprehensive abstracts. Our code and dataset can be accessed at \url{https://github.com/THU-KEG/TWAG}