Learning to Selectively Learn for Weakly Supervised Paraphrase Generation with Model-based Reinforcement Learning.

Haiyan Yin, Dingcheng Li, Ping Li

Browse the full NAACL paper archive.