BriefGPT.xyz
Aug, 2019
模仿学习与强化学习在改写生成中的实证比较
An Empirical Comparison on Imitation Learning and Reinforcement Learning for Paraphrase Generation
HTML
PDF
Wanyu Du, Yangfeng Ji
TL;DR
本研究通过pointer-generator文本生成模型的实验对比，表明在生成同义句时，模仿（IL）学习比强化（RL）学习更有效且优于目前的同类方法。
Abstract
Generating paraphrases from given sentences involves decoding words step by step from a large vocabulary. To learn a
decoder
,
supervised learning
which maximizes the likelihood of tokens always suffers from the e
→