Although the arrival was Siri Apple phones were a really big event and to some extent a revolution a few years ago, today it's an artificial assistant Applu viewed lightly through fingers. Siri it is limited in some respects compared to its competitors and therefore suffers slightly. However, this Apple Of course, he tries to fix it and improve his assistant to bring her back to the limelight. He then posts his progress in this regard on his blog from time to time. Machine Learning Journal, který se mimo jiné umělé inteligenci a strojovému učení využívaného u Siri And a new post has appeared on this blog, which describes how Apple seeks to improve search and name recognition for smaller local businesses in the US by using Siri.
You could be interested in
Pro Siri The names of some smaller businesses that he doesn't know can be a big challenge, which is why he makes mistakes from time to time. To reduce the error rate, he started Apple use for better name recognition Siri the user's geolocation, which, in very simple terms, helps to better understand the names of businesses. About all businesses in the vicinity Siri knows in advance, which gives him a certain head start in recognizing the name.
Name recognition using geolocation is supported by a total of 169 statistical areas in the US, which overlap in various ways and cover the vast majority of businesses and the population. If users do not have location services turned on, Siri is based on one global “file”. Thanks to all this, it should be Siri much more precise and although the entire blog post is really packed with a lot of technical terms, it can be simply said that it should give you Siri As a result, it will offer you the desired result much more often, in the sense that if you are in Kansas and want to visit Marty's Tavern, it will offer you a tavern with that name located in Kansas and not in Boston.
You could be interested in
The tool is currently only being used in the US, but according to a blog post, its expansion beyond the borders is quite likely. The entire system is, at least according to Applu is absolutely language-independent, so it should not be a problem to implement it practically anywhere.
