Google is upgrading its no-coding, in-browser AI trainer to make it more functional

Google has been upgraded Learnable machine, an in-browser web experiment that was originally released in 2017 to teach people some basic principles of machine learning.


The original version of the app was really just a nice little experiment that could be used to classify visual data from your webcam. (I learned to identify my house plants!) But Google has added new modes to the system, along with the option to export trained models, making Teachable Machine 2.0 a more functional system for building true AI tools.

Together with image data, Teachable Machine now works with audio and body pose input. Users can upload their own pre-collected data sets, sort data into more than three categories, and download and deploy their models locally or host in the cloud. That means that you can train a basic system with Teachable Machine and work on a website or app.

Google already offers a no-code called AI-trainer Cloud AutoML, but this is a much more professional tool, with more options for customization, scaling and customer support.

Teachable Machine 2.0 is fast and dirty in comparison: it is a driveway for new ML practitioners and something that users can use to quickly prototype an AI solution. Google notes that Teachable Machine runs entirely on the user's computer, which means that training data never leaves your device (reassurance for those who are concerned about privacy).

In the years since the last wave of machine learning systems started, there are numerous non-coding AI tools nearby. Experts are sometimes skeptical about the quality of AI models that they produce, and note that they can be inefficient and sloppy, and that, without the right programming skills, the people who build these tools don't really get the best out of them .

But it is undeniable that visual interfaces remove many of the intimidating factors of machine learning and can entice more people to experiment and tinker with these tools. The hard work of refinishing and refining a model can come later if necessary; it's just good to help people on their way. You can play with Teachable Machine yourself here.