Teachable Machine

What is this?

This experiment lets anyone explore how machine learning works, in a fun, hands-on way. You can teach a machine to using your camera, live in the browser – no coding required. You train a neural network locally on your device, without sending any images to a server. That’s how it responds so quickly to you. Watch this video to learn more:

What kind of things can I do?

Here are some links to things people have done so far: Make your hand say moo. Rock out by wiggling your fingers. And stay tuned, we’ll add more examples here soon. (Want to share something with us? Use the record button and share it on social media with #teachablemachine so we can check it out.)

Any tips I should keep in mind?

Capture at least 30 images per class. Be aware of when you’re pressing and releasing the button (that’s when it starts/stops capturing images). And you might need to capture lots of angles or variations of whatever it is you want your machine to recognize.

Why isn’t my machine working the way I want it to?

Don’t worry. Keep playing around. Seeing what works and what doesn’t is one way to explore how machine learning works. Keep in mind that your machine doesn’t have an understanding of higher level concepts, like faces or objects. It’s learning through the examples you give it. So if it’s not working the way you want, you might want to click the x to reset your classes and try out different approaches.

Where can I find more things like this?

Check out Wekinator by Rebecca Fiebrink, one of the inspirations for this project. It lets anyone use machine learning through simple actions instead of code. Here are some interactive guides for learning about machine learning. And check out other fun projects like this and this.

Are my images being stored on Google servers?

No. All the training is happening locally on your device.

How do I learn more about machine learning?

Here’s an intro-level video explainer. This site lets you interact with neural networks in more detail. And this free online course lets you dive in even deeper.

How was this built?

The image classification is powered by a neural network. It was made possible by Nikhil Thorat and Daniel Smilkov, the team behind TensorFlow.js. It’s an open-source library that allows web developers to train and run machine learning models locally in the browser. The code for this experiment is open-sourced here on Github.

We also made a boilerplate project which demonstrates how to use TensorFlow.js to create projects of your own like Teachable Machine here.

Who made this?

This experiment was a collaborative effort by friends from Støj, Use All Five and Creative Lab and PAIR teams at Google.