Google's ML Kit Lets Any App Put Machine Learning to Work
Google today announced ML Kit, a new SDK that will allow developers to add machine learning to their own applications. ML Kit gives developers access to five ready-to-use APIs: text recognition, face detection, barcode scanning, image labeling, and landmark recognition. Google believes these base APIs will let developers add a whole new level of smarts to their applications. ML Kit is available for both Android and iOS, and includes both on-device and in-the-cloud offerings. The image labeling API, for example, includes 430 sample images for offline, on-device object recognition, but more than 10,000 in the cloud. Developers will be able to add the offline component to their apps for free, though accessing the APIs via the cloud will carry nominal fees. Google said it worked with Qualcomm and Huawei, both of which have machine learning built into their most recent flagship processors, to generate its ML Kit models. The idea behind ML Kit is to bring machine learning to any device, regardless of the processor inside. "We want machine learning to be a thing," said Brahim Elbouchikhi, Product Manager at Google. Developers will be able to test different models side-by-side without doing any computation their own. Elbouchikhi envisions ML Kit will be most useful in camera apps, fitness apps, meal planning apps, fishing apps, and any other app where object recognition and labeling provides an extra layer of information and detail. "A fishing app [enhanced by ML Kit] can help you figure out what fish you just caught and estiamte the size," remarked Elbouchikhi. Developers can find ML Kit in Google's Firebase console.
Jun 5, 2017
Apple says iOS 11 will take advantage of machine learning to a much higher degree than before. The the company has new, specific APIs that developers can use in apps such as the camera for face tracking, face detection, landmarks, text, and so on.
Jan 25, 2018
Google today updated the Android Wear SDK for developers and revealed several new features that are headed to its smartwatch platform later this year. Primarily, Android Wear will add a notification indicator to all watch faces.
Apr 5, 2019
Google has turned its Smart Reply technology into an API that third-party Android and iOS developers can easily deploy in their own messaging apps. The software looks at the (up to) ten most recent messages in a thread, applies Google's machine learning algorithm locally (nothing is sent to the cloud,) and suggests three complete replies that can be sent with one tap.
May 23, 2018
Qualcomm today announced the Snapdragon 710, the first system-on-a-chip in its 700 series mobile platform. The 710 finds space between the Snapdragon 600 series and Snapdragon 800 series for devices that need to balance performance and cost.