A Diversity-Promoting Objective Function for Neural Conversation Models
該論文提出了基於最大互信息（Maximum Mutual Information—MMI）的期望計算思想。 傳統的生成模型計算方式是：target = argmax（logP（T|S）），即在source給定的條件下，尋找出最大概率的target。 這樣的計算方式所帶來的問題是會生成一般性的，具有大概率事件的target，失去了target的多樣性。 由互信息公式得： 該論文基於MMI思想，提出了
相關文章
- 1. A Diversity-Promoting Objective Function for Neural Conversation Models
- 2. A Diversity-Promoting Objective Function for Neural Conversation
- 3. A Diversity-Promoting Objective Function for Neural Conversation Models論文閱讀
- 4. "A Diversity-Promoting Objective Function for Neural Conversation Models"閱讀筆記
- 5. short text conversation: neural network
- 6. A Knowledge-Grounded Neural Conversation Model
- 7. lijiwei2-A Persona-Based Neural Conversation Model
- 8. A Knowledge-Grounded Neural Conversation Model 翻譯
- 9. Deep Reinforcement Learning for Dialogue Generation 論文閱讀 A Diversity-Promoting Objective Function fo
- 10. A Persona-Based Neural Conversation Model論文
- 更多相關文章...
- • Lua for 循環 - Lua 教程
- • SQL MIN() Function - SQL 教程
- • Java 8 Stream 教程
- • 爲了進字節跳動，我精選了29道Java經典算法題，帶詳細講解
相關標籤/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。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. A Diversity-Promoting Objective Function for Neural Conversation Models
- 2. A Diversity-Promoting Objective Function for Neural Conversation
- 3. A Diversity-Promoting Objective Function for Neural Conversation Models論文閱讀
- 4. "A Diversity-Promoting Objective Function for Neural Conversation Models"閱讀筆記
- 5. short text conversation: neural network
- 6. A Knowledge-Grounded Neural Conversation Model
- 7. lijiwei2-A Persona-Based Neural Conversation Model
- 8. A Knowledge-Grounded Neural Conversation Model 翻譯
- 9. Deep Reinforcement Learning for Dialogue Generation 論文閱讀 A Diversity-Promoting Objective Function fo
- 10. A Persona-Based Neural Conversation Model論文