Generating Related Work

TL;DR

This paper introduces a content planning and surface realization model for automatically generating related work sections by modeling citation motivation and outperforming existing summarization methods on an ACL dataset.

Contribution

It presents a novel content planning approach that constructs citation trees before surface realization, advancing automatic related work generation.

Findings

Outperforms state-of-the-art summarization models on related work generation

Introduces a new ACL Anthology dataset for related work generation

Models citation motivation to improve content relevance

Abstract

Communicating new research ideas involves highlighting similarities and differences with past work. Authors write fluent, often long sections to survey the distinction of a new paper with related work. In this work we model generating related work sections while being cognisant of the motivation behind citing papers. Our content planning model generates a tree of cited papers before a surface realization model lexicalizes this skeleton. Our model outperforms several strong state-of-the-art summarization and multi-document summarization models on generating related work on an ACL Anthology (AA) based dataset which we contribute.