Unpaired Sentiment-to-Sentiment Translation: A Cycled Reinforcement Learning Approach
Abstract
The goal of sentiment-to-sentiment "translation" is to change the underlying sentiment of a sentence while keeping its content. The main challenge is the lack of parallel data. To solve this problem, we propose a cycled reinforcement learning method that enables training on unpaired data by collaboration between a neutralization module and an emotionalization module. We evaluate our approach on two review datasets, Yelp and Amazon. Experimental results show that our approach significantly outperforms the state-of-the-art systems. Especially, the proposed method substantially improves the content preservation performance. The BLEU score is improved from 1.64 to 22.46 and from 0.56 to 14.06 on the two datasets, respectively.
Keywords
Cite
@article{arxiv.1805.05181,
title = {Unpaired Sentiment-to-Sentiment Translation: A Cycled Reinforcement Learning Approach},
author = {Jingjing Xu and Xu Sun and Qi Zeng and Xuancheng Ren and Xiaodong Zhang and Houfeng Wang and Wenjie Li},
journal= {arXiv preprint arXiv:1805.05181},
year = {2018}
}
Comments
Accepted by ACL 2018