Mobile Keyboard for Android Leverages Neural Networks

The six-year-old SwiftKey Keyboard for Android has been re-architected to leverage cutting-edge neural networking technology to provide improved predictive typing options, TouchType Ltd. announced yesterday.

The new technology has been infused into the SwiftKey Keyboard app as a Google Play update.

"With this update SwiftKey is now able to meaningfully capture the relationship and similarity between words," the company said in a blog post yesterday. "For example, having previously seen the phrase 'Let's meet at the airport,' the technology is able to infer that 'office' or 'hotel' are similar words which could also be appropriate predictions in place of 'airport.'

"Further, it understands that 'Let's meet at the airport' has a similar sentence structure to 'Let's chat at the office.' This intelligence allows SwiftKey to offer you the most appropriate prediction or autocorrection based on the sentence being typed."

The neural network technology was first used in an early experimental prototype, announced last October. The company described it as "the first instance of neural networks being used locally on a smartphone."

TouchType said the integration of neural networking technology into its keyboard paves the way for more future changes and improvements.

In the meantime, the update is first being rolled out to U.S. and U.K. English language models, with more language support planned.

Here's an infographic with more details on how neural networks improve typing predictions:

How Do Neural Networks Make Better Word Predictions?
[Click on image for larger view.] How Do Neural Networks Make Better Word Predictions? (source: TouchType)

About the Author

David Ramel is an editor and writer for Converge360.