SentiGAN: Generating Sentimental Texts via Mixture Adversarial Networks

NLPIR SEMINAR 20th ISSUE COMPLETED

Last Monday, Ziyu Liu gave a presentation about the paper, SentiGAN: Generating Sentimental Texts via Mixture Adversarial Networks, and shared some opinion on it.

The evaluation metrics to measure the quality of generated sentences contains four aspects – fluency, novelty, diversity and intelligibility. Through the case study, the sentences produced by SentiGAN are more readable and of higher quality.

此条目发表在Seminar研讨班, Update动态分类目录。将固定链接加入收藏夹。

发表回复