该论文提出了基于最大互信息（Maximum Mutual Information—MMI）的期望计算思想。传统的生成模型计算方式是：targetargmax&#x

本文是李纪为的论文“A Diversity-Promoting Objective Function for Neural Conversation Models”阅读笔记。违章提出使用MMI代替原始的maximum likelihood作

本篇分享的文章是前一篇分享A Persona-Based Neural Conversation Model的pre-paper，题目是A Diversity-Promoting Objective Function

A Diversity-Promoting Objective Function for Neural Conversation Models SEQ2SEQ模型用于conversational responses倾向于产生safe&am