Jointly Optimizing Diversity and Relevance in Neural Response Generation

NAACL 2019 Xiang GaoSungjin LeeYizhe ZhangChris BrockettMichel GalleyJianfeng GaoBill Dolan

Although recent neural conversation models have shown great potential, they often generate bland and generic responses. While various approaches have been explored to diversify the output of the conversation model, the improvement often comes at the cost of decreased relevance... (read more)

PDF Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Dialogue Generation Reddit (multi-ref) SpaceFusion relevance (human) 2.72 # 1
interest (human) 2.53 # 1