One day, your phone's keyboard will know and understand you so well it'll predict entire text messages in your tone of voice. But until then, SwiftKey Neural Keyboard promises more accurate and useful next-word predictions.
"We're using neural networks to enable our keyboards to understand language at a deeper level," the company wrote in a blog post. "We've rebuilt SwiftKey's language engine from the ground up using the power of neural network technology—the first instance of neural networks being used locally on a smartphone."
Android owners can now download the app, initially available in US English and UK English, with more languages to come.
Based on an artificial intelligence platform inspired by the human brain, this update boasts a better understanding of sentence construction for a more "human" touch.
The keyboard understands word similarity, allowing the program to predict words that may not have been paired during the machine-learning phase. For example, having encountered the phrase "Let's meet at the airport" during training, the technology can infer that "office" or "hotel" are similar words that could be appropriate replacements for "airport."
Swift says this is first use of neural network language models on mobile; it's typically reserved for larger servers that require significant computational resources. "Re-architecting the keyboard to use neural networks lays the groundwork for many exciting changes and improvements to come," the blog post teased.