Teachable Machine by Google is a web-based tool that democratizes machine learning by making it fast, easy, and accessible to all. It provides a user-friendly interface to train a computer to recognize images, sounds, and poses without the need for coding or machine learning expertise.
Easy Model Creation
With Teachable Machine, you can create machine learning models in three simple steps:
- Gather: Collect and group examples into classes that you want the computer to learn.
- Train: Train your model and instantly test it to assess its classification accuracy.
- Export: Export your model for use in your projects, whether they’re websites, apps, or more.
Teachable Machine is flexible and accommodates different work styles. It allows you to:
- Use files or capture examples live.
- Teach a model to classify images using files or your webcam.
- Teach a model to classify audio by recording short sound samples.
- Teach a model to classify body positions using files or by striking poses in your webcam.
The tool operates on-device, ensuring that no webcam or microphone data leaves your computer.
Users have creatively applied Teachable Machine in various projects:
- Tiny Sorter: A DIY experiment connecting Arduino and Teachable Machine.
- Project Euphonia: Steve Saling uses Teachable Machine to communicate using facial gestures to trigger sounds.
- Teachable Snake: Vince MingPu Shao used Teachable Machine to turn their webcam and a piece of paper into a game controller.
Limitations and Concerns
Teachable Machine is a powerful tool, but there may be situations where your model doesn’t work as expected. This could be due to changing environments, different microphones, or bias in data. The tool is designed to help users understand these machine learning challenges and learn how to improve their models.
Potential Future Developments
Teachable Machine currently supports training with images, sounds, and poses. There is potential for the tool to support more types of training in the future, expanding its applicability across more fields and use cases.