Sequential Copying Networks

TL;DR

This paper introduces Sequential Copying Networks (SeqCopyNet), a novel model that enhances text generation by copying sequences of words from source sentences, improving performance in summarization and question generation tasks.

Contribution

The paper proposes SeqCopyNet, a new copying framework that copies sequences of words using pointer networks, extending beyond single-word copying in sequence-to-sequence models.

Findings

SeqCopyNet effectively copies meaningful spans from source sentences.

It outperforms baseline models in summarization and question generation.

The model improves the quality of generated text by leveraging sequential copying.

Abstract

Copying mechanism shows effectiveness in sequence-to-sequence based neural network models for text generation tasks, such as abstractive sentence summarization and question generation. However, existing works on modeling copying or pointing mechanism only considers single word copying from the source sentences. In this paper, we propose a novel copying framework, named Sequential Copying Networks (SeqCopyNet), which not only learns to copy single words, but also copies sequences from the input sentence. It leverages the pointer networks to explicitly select a sub-span from the source side to target side, and integrates this sequential copying mechanism to the generation process in the encoder-decoder paradigm. Experiments on abstractive sentence summarization and question generation tasks show that the proposed SeqCopyNet can copy meaningful spans and outperforms the baseline models.