One of Swiftkey's most famous keyboard apps for Android, it stands out for its accuracy in suggesting and correcting words when the user is typing, but it's good to get even better. This week the Swiftkey Neural app was launched. The technology built into it is able to significantly capture the relationship between the words typed by the user. Analyzing the context he understands them, and with this the neural model allows a better system of predicting terms.
For those who don't know, the current Swiftkey – the new other standalone app – uses another technology they call n-gram. She analyzes the last two words typed, following their reasoning he suggests the next. As for the neural network, it works similarly, but instead of two words technology parses all words and understands entire contexts, thus facilitating suggestions.
Another interesting point is that this is all done offline, without relying on a server to do the word analysis. According to Joe Braidwood, chief marketing officer at SwiftKey, this was not easy "because of the limitations of today's hardware." Believe if you want.
This is still an alpha version of the app. Although they had difficulty with the hardware, SwiftKey Neural uses the smartphone's GPU to be able to do the app calculations without requiring too much processor. Thanks to that, it runs well even on smartphones with medium specifications.
SwiftKey Neural Alpha
Did you like the app? Do not be so excited, as it is still an alpha version, a lot has yet to be improved in the neural Swiftkey and in addition it is only compatible with the English language for now. But for those who write a lot in English, it is worth testing, especially being free.
And you, what keyboard are you using? Although Swiftkey is a great app, I can't particularly let go of the Google keyboard.
. (tagsToTranslate) Swiftkey (t) android keyboard (t) neural swiftkey (t) android apps