After the screen is locked, make a phone call with siri and prompt to unlock it first. Why not unlock it and call directly?

Generally speaking, the first time you use siri to make a phone call, you need to unlock it, and then there is no need to use it again. However, it can be changed in the settings. Allow SIRI to use under lock screen.

Pay attention to the permissions of the software when locking the screen.

Siri is an intelligent voice control function applied by Apple on its iPhone4S, iPad 3 and above mobile phones and Macs. Siri can transform iPhone4S and above (tablets above iPad 3) into intelligent robots;

Through Siri, users can read short messages, introduce restaurants, ask about the weather, and set alarm clocks by voice.

SIRI is voice interpretation &; The acronym of recognition interface means speech recognition interface.

summary

Many people will answer that Siri uses artificial intelligence and cloud computing. Indeed, in general, it is these two technologies. However, this overview feels almost meaningless. Why not just say "computing technology" (note that it is not computer technology)?

So in this article, I will introduce the technology that I know Siri may adopt (not necessarily accurate due to personal guess).

Front end aspect

At the front end, that is, user-oriented and user interface (UI) technology, mainly speech recognition and speech synthesis technology.

Speech recognition technology is to convert users' spoken language into words, which requires a strong voice knowledge base, so the so-called "cloud computing" technology is needed.

Speech synthesis is to convert the returned text results into speech output. This technology can theoretically be done locally, but I don't know if Siri is like this. Of course, it doesn't have to be done in the cloud. Under the current wireless bandwidth, voice traffic is nothing at all.

Extended data:

Knowledge computing

(Search) Technology (Computational Knowledge)

The representative of this technology is WolframAlpha.

Unlike searching Internet information, Wolfram|Alpha will mine and build an extremely large organized database from the public (including public web pages, etc.). ) and authorized resources;

Then it is processed by advanced natural language algorithm, and finally a tool similar to Google search is constructed.

Unlike web search technology, in this system, the answer is highly structured. For example, if you search China, you can get all kinds of parameters and materials related to China, and present them in a form close to a table.

Wolfram|Alpha can also understand some natural languages, such as how old are you? This will answer Wolfram|Alpha's age. If you want to test this technology, please visit Wolfram|Alpha.

This technology should be available in China, but there is no product yet. Maybe Baidu's box computing is half. So there is a great chance.

Knowledge base technology

Compared with web search technology, it is basically based on a term or topic, so the data obtained is of high value, large amount of knowledge and good structure.

Compared with knowledge computing technology, these technologies need human participation, which has both advantages and disadvantages. The advantage is that, after all, people are smarter than machines for the time being, and the edited knowledge is richer and more accurate;

The disadvantage is that manpower is limited, even if the power of the community is mobilized like Wikipedia, it can't produce enough knowledge, and knowledge calculation can only produce "infinite" knowledge theoretically with enough algorithms.

Question and answer recommendation technology

In fact, this can't be called technology, it should belong to the technology of knowledge base. The difference is that this technology is aimed at some life information with a high degree of localization, and the typical representative is Yelp. Because this thing is relatively simple, I won't introduce it in detail.

Language model technology

This is how to improve Siri's ability to recognize the names of local interest points such as enterprises and restaurants, and integrate the knowledge of users' location, which is called Geo-LM, so as to improve Siri's voice recognition ability. These models allow Siri to better evaluate the user's pronunciation content.

References:

Siri- Baidu encyclopedia