Joint Copying and Restricted Generation for Paraphrase

TL;DR

This paper introduces a novel sequence-to-sequence model that combines copying and restricted generation decoders for improved paraphrase generation, outperforming existing methods in informativeness and language quality.

Contribution

A new Seq2Seq model integrating copying and restricted generation decoders with a mode predictor for enhanced paraphrase generation.

Findings

Outperforms state-of-the-art approaches in paraphrase quality

Effective combination of copying and rewriting modes

Improves informativeness and language quality in generated paraphrases

Abstract

Many natural language generation tasks, such as abstractive summarization and text simplification, are paraphrase-orientated. In these tasks, copying and rewriting are two main writing modes. Most previous sequence-to-sequence (Seq2Seq) models use a single decoder and neglect this fact. In this paper, we develop a novel Seq2Seq model to fuse a copying decoder and a restricted generative decoder. The copying decoder finds the position to be copied based on a typical attention model. The generative decoder produces words limited in the source-specific vocabulary. To combine the two decoders and determine the final output, we develop a predictor to predict the mode of copying or rewriting. This predictor can be guided by the actual writing mode in the training data. We conduct extensive experiments on two different paraphrase datasets. The result shows that our model outperforms the state-of-the-art approaches in terms of both informativeness and language quality.