Google is always hard at work on a slew of projects related to all things tech, committed to exploring the boundaries of technology for a more interconnected, tech-literate world. One of their latest projects is Teachable Machine, a neural network-based experiment that helps people learn about the mechanics of machine learning right from a browser.

The website is part of the Google domain Experiments, an online space where coders can host their projects made using AI, AR, Android, Chrome, and other tools. Established in 2009, AI Experiments allows developers, engineers, and other makers to collaborate and publicly share their creations online.

Teachable Machine was created by design studios Stoj, Use All Five, as well as Google’s Creative Lab and PAIR teams. The project was partly inspired by Wekinator, a popular tool for creative machine learning applications built by Rebecca Fiebrink, professor at Goldsmiths, University of London. Fiebrink is also the instructor for our course Machine Learning For Musicians and Artists.

Through simple physical gestures, the experiment offers a fun, kinetic way to learn machine learning online without any coding. Information for Teachable Machine is stored in your own device instead of the Google server, and its easy-to-use interface is suitable for anyone who wants to learn more about AI. Users are encouraged to get creative; whether you wave your hand or make a silly face, the computer will try to learn your unique actions to choose a specified output, such as a GIF or sound.

Try the Teachable Machine for yourself! For a more in-depth look at a similar system, learn about Wekinator here:

Machine Learning for Musicians and Artists