The Apple Worldwide Developer Conference (WWDC) 2023 keynote concluded on June 5 with a slew of announcements, including features in the realm of artificial intelligence (AI).
Among the hardware unveiled was the long-awaited 15-inch MacBook Air, Mac Studio, and Mac Pro with the M2 chipset. Software releases include the latest iOS 17, iPadOS 17 and watchOS 10, all of which were to be announced at the event.
Of course, the center of attention was Apple’s augmented reality (AR) headset called Vision Pro.
“Welcome to the era of spatial computing with Apple Vision Pro. Never seen anything like this before!” Apple CEO Tim Cook tweeted along with a video showcasing the device’s amazing capabilities.
But while the Vision Pro is clearly the spotlight, Apple also introduced AI-based features to several of its products at WWDC 2023.
Interestingly, though, given that competitors Google and Microsoft are well ahead of Apple in that space, despite being the most talked-about word at the moment, Apple did not appear in the presentation. You didn’t mention “AI” once.
Apple’s term was “machine learning” (ML). Machine learning is commonly understood as another term for AI, but is technically a subset of the former.
Microsoft defines machine learning as “the process of using mathematical models of data to enable computers to learn without direct direction.” This means that computer systems can learn from their own experience and improvise.
Nonetheless, it’s worth noting that Apple emphasized machine learning in some of its products as it mentioned a “transformer language model” during the presentation.
For the uninitiated, a transformer language model is simply a neural network that can learn context from input and respond in a consistent manner. Google first defined this in his 2017 paper. The Transformer architecture is behind some of the most famous AI chatbots such as ChatGPT.
Features Apple talked about machine learning
Improved AutoCorrect and Dictionary in iOS17
Craig Federighi, Apple’s SVP of Software Engineering, mentions both machine learning and the Transformer language model when talking about the new Dictation and AutoCorrect features during the iOS 17 presentation at WWDC 2023. Did.
“AutoCorrect utilizes on-device machine learning, and we have continued to evolve these models over the years. , AutoCorrect is more accurate than ever,” he said.
Federighi added that Apple Silicon will allow the iPhone to run this model every time the user taps a key.
In a statement, Apple said the Transformers language model improves user experience and accuracy every time you type. The company says users will see inline predictive text suggestions as they type, and can insert complete sentences by tapping the spacebar.
The autocorrect feature also works at the sentence level, allowing the machine to suggest whole sentences as messages are typed. The feature will also learn about your typing habits, such as intentional misspellings and “ducking words,” meaning swear words, and will be able to suggest words and phrases accordingly.
Federighi also said that the dictation feature uses a “transformer-based speech recognition model powered by a neural engine” to improve accuracy.
Journal — a new app for journaling in iOS 17
Shortly after Federighi’s announcement of AutoCorrect and Dictionary features, Adeeti Ullal, senior manager of sensing and connectivity at Apple, introduced a new app called Journal.
The app was one of the most talked about features after the event. It is essentially a diary app that allows users to record their moments and memories as a means of improving their health.
Scheduled to launch in September 2023, Apple said in a statement that the app would use “on-device machine learning” to make personalized suggestions.
“Suggestions are intelligently curated from information on your iPhone such as photos, places, music, and workouts,” Ural said.
Apple said even the company cannot access users’ Journal entries, ensuring the highest level of privacy and end-to-end encryption.
Developers will be able to add journaling suggestions to their apps using the app’s API.
Personas in Vision Pro
Personas is a unique feature that Apple introduced in Vision Pro at WWDC 2023. The AR device’s camera can scan the user’s face and create a lifelike caricature that can be projected onto the other user during a video call or his FaceTime interaction.
It’s like a digital clone of a real user, allowing other participants in the video conference to make the same gestures and facial expressions that a real user does without having a Vision Pro overhead. You can see your avatar.
In its presentation, Apple said personas were “created using Apple’s cutting-edge ML technology.”
Other features using Apple machine learning
Apart from iOS 17 features and the new Journal app, Apple also mentioned machine learning when talking about PDF features in iPadOS, lock screens on iPad devices, and smart stack widgets on Apple Watch.
The Photos app can now use an improved ML program to identify a user’s pet, distinguishing a user’s cat or dog from other cats and dogs.
Apple announced that the AirPods Pro’s adaptive audio feature “uses ML to understand environmental conditions and listening preferences over time.” It can also sense the user’s behavior and environment and automatically adjust the volume or turn off the noise canceling function.
(Hero and Featured Images: Apple)
