Google’s AIY Projects platform aims to make AI development quick, easy, and accessible for everyone
Google is releasing a set of new kits for their AIY Projects platform, and these all-in-one packages, the Voice Kit and the Vision Kit, are designed to make it easy to create your very own AI-driven device.
As one of the leaders in the sector, Google is investing heavily in all things relating to artificial intelligence, and according to their Google Developers blog, the updated kits make it so that STEM students (and hobbyists) can easily create a Google Assistant smart speaker or object-recognizing smart camera.
Google is working in a lot of areas to make machine learning more accessible, and the original AIY Projects (as in, DIY and AI, get it?) were introduced last year, but they weren’t really suitable for newcomers as they shortsightedly lacked a few key components, requiring users to already have their own spare Raspberry Pi, for example.
However, the new kits come with everything out of the box, including a fresh Raspberry Pi Zero and a preloaded SD card, and are supported by a stack of companion materials such as a dedicated app and a revamped website with clear and concise instructions.
The Voice Kit
The Voice Kit is an all-in-one package that lets budding innovators create a voice-controlled speaker unit that supports Google Assistant.
Some might think that the kit isn’t the prettiest as it’s constructed from cardboard, but that’s not really what’s important here as it does offer full voice assistance voicetech capabilities.
In fact, the Voice Kit and comes loaded with all the necessary wires, circuitry, speaker and housing needed to create your personal voice assistant.
The Vision Kit
The Vision Kit features the same cardboard housing as the Voice Kit, but comes with a built-in “intelligent” camera that’s capable of identifying faces, objects, and can even recognize emotions. It’s compatible with a variety of image recognition neural networks, and comes with the aforementioned Raspberry Pi Cam 2 camera.
The standard test, once the Vision Kit has been put together, is called “Joy Detector,” which is where you point the camera and somebody’s face and get them to smile, then again but with a really big smile, and then finally with a frowny face.
The machine learning software should then be able to detect whether or not that person is smiling or frowning.
Users will need to assemble all the components of each kit themselves, but comprehensive guides are available online, and the app is also there to help set up and configure the kit. Currently the app is only available via Android, but Google has said that apps for iOS and Chrome are in the pipeline.
Both kits are now available online and in Target stores across the U.S., with plans for more availability internationally as time goes on. The kits are priced at $90 for the Vision Kit, and $50 for the Voice Kit, which is somewhat of a price hike from the previous AIY Project packs, but does include all that juicy hardware.
The hope is that the kits are now better suited to STEM students looking to delve into the world of AI and machine learning, and Google has also revamped TensorFlow with TensorFlow.js, a browser-based version of their industry-leading machine learning framework.
So conceivably, we mere mortals now have a good chance at training our own smart camera and voice assistants without needing a PhD. That sounds and looks like a great idea.