Does your phone instantly recognise your face? Have you used apps that know what song is playing? If so, you might know more about machine learning than you think! These systems aren’t difficult to build, especially with tools like Google’s Teachable Machine that learn from your webcam, microphone, or uploaded files. You don’t need to know how to code either.
So, if you’re looking to build a machine learning model to do something like…
- Identifying your cat at the cat flap
- Build a pose classifier for your yoga routine
Or even…
- Build a conservation system that can identify different individuals of a species
- Build a system to help diagnose different skin conditions
Here are the basics to get started
Getting started takes just three easy steps:
- Gather Data: The more data you add, the more accurate the outcome will be! You’ll need to create categories with your data (like ‘dog’ vs ‘not a dog’). Your data can be audio or images; you choose!
- Train: Click ‘Train Model’ and Teachable Machine will process the data in seconds.
- Export: Test it live, then export your AI model to use wherever you would like!
Our Top Tips 👀
- Mix things up by feeding your model photos in different lighting, from a variety of angles, and against different backgrounds, so it learns accurately.
- Keep your background simple for a lot of the learning process, so the machine focuses on what you’d like it to.
- Add at least 30 to 50 samples per category for the best results!
