Apple’s future explained — the iPhone 13, 13 Pro and iPad machine learning future
Today was Apple’s mobile device update, where it unveiled the latest iPhone, Apple Watch, and iPad models. The most prominent technology featured across all devices was the company’s machine learning A13 and A15 Bionic chips. It makes sense that the future of Apple is in machine learning since it depends on a chip that only Apple has the capability to produce (so far). This has the added advantage of making the company’s suite of hardware products more defendable and integration-friendly.
New iPad — First up, the new iPad comes with Apple’s A13 Bionic chip which has a 20% faster CPU and GPU, with a neural engine that is 3x faster than the best-selling Chromebook and 6x faster than the best selling Android tablet (according to Apple). Apple designed this neural engine to power its next-gen machine learning capabilities, like for the Live Text feature that uses on-device intelligence to read and process text in a photo. This could be used to pull text like an email address from a picture of a handwritten note. Other applications could include snapping a photo of a storefront and tapping the photo to call a number directly from a sign in the store window.
The iPhone 13 — Next, the iPhone 13 features the new A15 Bionic. Built with 5nm processors, this CPU is 50% faster. However, the iPhone 13’s new 16-core neural engine is the main event, capable of 15.8 trillion operations per second to enable faster machine learning computations. A few developer use-cases include apps like SwingVision, which runs augmented reality for real-time shot tracking, video analysis, and remote coaching; Peakvisor, which identifies mountains and allows users to explore detail-rich maps; and Seek, which processes millions of photos to instantly identify plants and animals around you.
Cinematic Mode — Apple also uses this neural engine to power its new video “cinematic mode” which adds a rack-focus style shooting capability to automatically adjust focus. This is used to hold focus on a subject or create focus transitions in real-time by anticipating when a subject enters the frame or looks away from the camera.
The 13 Pro’s cinematic mode uses machine learning to allow a video’s depth of field to be edited in post-production.
What’s to come — New machine learning features can be rolled out with software updates. We expect that Apple will continue to release new features that depend on neural networks and machine learning in the coming months — features like automatic photo object detection, machine learning APIs for app developers, video analyzing tools to identify emotions, and so on. If Apple chips have multiple CPUs optimized and devoted to machine learning, the future of predictive analytics, medical diagnosis, emotion recognition, preventative insights, and much more is just around the corner.