CRPO: Confidence-Reward Driven Preference Optimization for Machine Translation.

Guofeng Cui, Pichao Wang, Yang Liu, Zemian Ke, Zhu Liu, Vimal Bhat

Browse the full ACL paper archive.