知识卡片
候选生成与回应选择应分别定义
解析提取信息,分析形成特征和分数,生成产生候选,决策按对话目标选择回应。把输出再送入分析模块可比较候选质量,但评分目标需要独立设计;语言流畅或分数高,不自动意味着真实、相关与有帮助。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- · 查看原书依据
Parse —Extract features, structured numerical data, from natural language text. Analyze —Generate and combine features by scoring text for sentiment, grammaticality, and semantics. Generate —Compose possible responses using templates, search, and language models. Decide —Decide which generated response is most likely to move the conversation closer to the user’s conversational goal. Database —Store conversation history, user information, and general world knowledge for use in the
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第01章《会读写的机器与自然语言处理》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-1.html)
- 结论依据:摘录中关于候选生成与回应选择应分别定义的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:Parse —Extract features, structured numerical data, from natural language text. Analyze —Generate and combine features by scoring text for sentiment, grammaticality, and semantics. Generate —Compose possible responses using templates, search, and language models. Decide —Decide which generated response is most likely to move the conversation closer to the user’s conversational goal. Database —Store conversation history, user information, and general world knowledge for use in the