SwiftKey for Android includes a built-in neural network to help it make better predictions and autocorrections.
(CCM) — SwiftKey has added a neural network to its SwiftKey Keyboard for Android app to enable it to better capture the relationship and similarity between words. This helps the app to make sense of sentences and is intended to make it better at predicting the next word a user plans to type or at autocorrecting a word as it is typed, the company said in a blog post. "Having previously seen the phrase 'Let’s meet at the airport,' the technology is able to infer that 'office' or 'hotel' are similar words [that] could also be appropriate predictions in place of 'airport,' it explained.
Neural networks are modeled on the way the human brain works and need training or "tuning" to help them carry out tasks without being given formal instructions or rules to follow. SwiftKey claims that its app is the first instance of a neural network being used locally on a smartphone, rather than running on a more powerful computer system in the cloud. The neural network is being rolled out in the U.S. and U.K English versions of the app, and current users will be upgraded automatically. SwiftKey was acquired by Microsoft last February in a deal thought to be worth $250 million.
Image: © SwiftKey.