Joint Copying and Restricted Generation for Paraphrase

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.

7 pages, 1 figure, AAAI-17

Cited by in corpus (8)

• Unsupervised Paraphrase Generation using Pre-trained Language Models
• Zero-Shot Paraphrase Generation with Multilingual Language Models
• Paraphrase Augmented Task-Oriented Dialog Generation
• Unsupervised Contextual Paraphrase Generation using Lexical Control and Reinforcement Learning
• Towards Diverse Paraphrase Generation Using Multi-Class Wasserstein GAN
• Generating Syntactically Controlled Paraphrases without Using Annotated Parallel Pairs
• PowerTransformer: Unsupervised Controllable Revision for Biased Language Correction
• Decomposable Neural Paraphrase Generation