lijiwei2-A Persona-Based Neural Conversation Model
使用person embedding embedding 的作用 蕴含背景和说话方式，解决了多轮对话中response不一致的问题。（换一种方式问问题，得到的答案不同）。 embedding的实现 具体的人物做背景标注昂贵，根据对话的回答聚类产生。 在target的E表示中加入speaker representation。每个source用户绑定一个向量vi，在decode的时候和h_t,e_t（
相关文章
- 1. A Knowledge-Grounded Neural Conversation Model
- 2. A Persona-Based Neural Conversation Model论文
- 3. A Knowledge-Grounded Neural Conversation Model 翻译
- 4. short text conversation: neural network
- 5. [dialog system]阅读笔记 - 2016__A Persona-Based Neural Conversation Model
- 6. A Persona-Based Neural Conversation Model论文学习零散记录
- 7. A Diversity-Promoting Objective Function for Neural Conversation
- 8. A Diversity-Promoting Objective Function for Neural Conversation Models
- 9. A Neural Conversational Model
- 10. neural topic model
- 更多相关文章...
- • Thymeleaf条件判断 - Thymeleaf 教程
- • ionic 复选框 - ionic 教程
- • 常用的分布式事务解决方案
相关标签/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. A Knowledge-Grounded Neural Conversation Model
- 2. A Persona-Based Neural Conversation Model论文
- 3. A Knowledge-Grounded Neural Conversation Model 翻译
- 4. short text conversation: neural network
- 5. [dialog system]阅读笔记 - 2016__A Persona-Based Neural Conversation Model
- 6. A Persona-Based Neural Conversation Model论文学习零散记录
- 7. A Diversity-Promoting Objective Function for Neural Conversation
- 8. A Diversity-Promoting Objective Function for Neural Conversation Models
- 9. A Neural Conversational Model
- 10. neural topic model