Page 1
A smart speaker can answer questions, play songs, or set a timer after hearing just a few spoken words.
智能音箱听到几句话后，就可能帮我们放音乐、设提醒，或者回答问题。正文预览
Page 2
It does not understand speech the same way a person does, but it follows a chain of listening, matching, and responding.
它并不像人一样真正理解语言，而是通过听取、匹配和回应这一整套流程来完成任务。
Page 3
First, microphones pick up sound waves and turn them into signals a computer system can process.
首先，麦克风会接收声音，把声音变成电脑系统能处理的信号。
Page 4
The device then tries to recognize words by comparing the sound pattern to many examples it has learned from.
接着，设备会把听到的声音模式和它学过的大量语音样本进行比较，尽量识别出词语。
Page 5
After that, it looks for the intention behind the words, such as asking for weather, music, or a home control action.
识别出词语后，它还要继续判断这句话想表达的任务，比如查天气、放音乐，还是控制家电。
