I’m sure many Android users are jealous with iOs users because they are well supported by the personal assistant called Siri. Today we will show you how this personal assistant works!
Siri: The process of voice recognition and answering
First of all, Siri uses the Automatic speech recognition (ASR) to transcribe the speech of the owners into text. ASR technology is only capable of detecting short commands, dictations or questions. Mispronounced words also can’t be detected. Then, this assistant uses a natural language processing program to analyze the parts of speech, the phrases, the constituent parsing and dependency parsing of the text. The result comes out as something they call “parsed text”.
Secondly, Siri runs a question and intent analyzing program to study the parsed text, in order to detect user’s intentions, whether it’s a command, a question, or something else. Then, this assistant sends those data to the server in the cloud to find the solution. If the command can be handled locally in the device, the server will do nothing but recording the speech for further study. If external help is required to fulfill the command, the server will send those data to a 3rd-party website for better response.
Thirdly, when the answer is available, Siri will transform those data into written text in the same language as the input speech. After that, Siri runs text-to-speech technologies to transform the text into spoken language, which is the final result the user receives. It seems to be a complicated and long process for this assistant to finally come up with an answer. But in fact, these steps are quickly conducted and the answer often arrives after only a few seconds.
Siri: What’s next?
Currently, Siri is doing a very good job in assisting iPhone users. Billions of commands are fulfilled every week. This technology from Apple is no doubt the best personal assistant nowadays. According to many sources, this virtual assistant will be developed into a bigger system that can take control of the social media and smart devices.