Apple Exec Explains Inner Workings of the iPad’s Handwriting Recognition


As reported by Popular Mechanics, iPads running on iPadOS 14 can very accurately recognize anything you write or draw by hand using an Apple Pencil, and convert it into typed text using Scribble. Apple executive Craig Federighi shared more details about the feature with the publication.

The recently released iPadOS 14 makes use of machine learning to identify and interpret handwriting (and even drawing – shapes you doodle by hand in the Notes app will now be straightened out by your iPad). However, Apple’s approach to machine learning is unlike any other.

Apple’s strategy involves feeding its software a smarter, more dynamic, and more varied data set for human handwriting and its nuances, and then putting the iPad’s unprecedented processing power at its disposal.

“When it comes to understanding [handwriting] strokes, we do data-gathering. We find people all over the world, and have them write things,” explained Apple’s senior VP of software engineering Craig Federighi. “We give them a Pencil, and we have them write fast, we have them write slow, write at a tilt. All of this variation.”

Apple’s methodology allows its software to learn not how to interpret what humans write, but to learn how they write. This, combined with knowledge of a language’s syntax, allows iPads to determine what character or symbol you’re going to write next and what word you’re going for as a whole.

And unlike systems like Alexa and Siri that outsource their data processing to the data centers they are connected to, your iPad does all of this data processing, statistical analysis, and handwriting comprehension locally on its own.

Apple’s remarkably accurate and eerily intuitive handwriting recognition is available on all iPad models compatible with iPadOS 14.