Google has added support for iPhones and iPads to its TensorFlow neural network. The artificial intelligence program has been a key aspect of the company’s recent advancements in the field. TensorFlow was instrumental in AlphaGo’s victory over Lee Seedol, and has been part of other Google initiatives as well. Like other neural network software, TensorFlow also aims to give computers the ability to compute data like the human brain does.
TensorFlow already works on Android, and its inclusion for iOS means more apps on the platform will be able to take advantage of the software. That said, don’t expect this to make your iPhone smarter just yet. While Google has added support for iOS in TensorFlow 0.9, it is currently unclear when a final version will be available for iOS.
Google’s goal with this is of course, to get more users feeding data to TensorFlow. Being an AI algorithm, it is designed to get smarter as more people use it. This means if developers make iOS apps using TensorFlow, it gives the AI more experience in the real world. While Android allows reach to the major portion of the market, Google couldn’t ignore iOS, which is the second largest. Google has also released TensorFlow as an open source system, allowing anybody to modify it, and hence add to it.
A TensorFlow powered app on iOS could perform intelligent functions, like analysing photos to recognise particular subjects, making better suggestions based on your usage, adding smart reminders to a task managing app, and more.
AI has been at the forefront of developments in the tech world recently. Companies like Google, Movidius, Microsoft and many others, have published bots and other programs that exhibit human-like intelligence, to some extent. Movidius recently announced its Fathom compute stick, which adds AI-based computer vision to other machines, while Google announced its Tensor Processing Units, a special ASIC designed for AI, predominantly at the server level.