知识卡片
语音识别与文本处理分开验收
语音助手通常需先将音频转换成文字,再进行意图、检索或生成,最后可将回应转成语音。本书主要处理转录后的文本。识别错误会影响下游语义,文本模型本身表现好也不能证明整条语音链路可靠。
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
This book focuses entirely on English text documents and messages, not spoken statements. Chapter 7 does give you a brief foray into processing audio files, using the example of Morse code. But apart from that, we focus on the words that have been put to paper … or at least put to transistors in a computer. There are whole books on speech recognition as well
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:摘录中关于语音识别与文本处理分开验收的定义、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:This book focuses entirely on English text documents and messages, not spoken statements. Chapter 7 does give you a brief foray into processing audio files, using the example of Morse code. But apart from that, we focus on the words that have been put to paper … or at least put to transistors in a computer. There are whole books on speech recognition as well