Why I built Perceptron Lab
I built this library to work through machine learning for myself. Perceptron Lab follows the same approach: start with the problem people were trying to solve, then work through why their solution makes sense.
Being handed a formula and shown that it works leaves a lot unexplained. I believe the history and motivation behind these methods make them comprehensible. We will work through their pieces before putting them together, with examples you can change and calculations you can inspect.
Here is what you get on Perceptron Lab.
History and motivation
Every lesson is grounded in the history of the mathematics, the logic, and the needs of the time.
Charts that are real models
Change the data and the model fits again, so you can test what a method does instead of taking my word for it.
Accessible to everyone
You don't need a degree or a strong background in mathematics to make progress. The site teaches the mathematics you need.
Comprehensive quizzes
Questions throughout each lesson, so you can feel your knowledge growing.
Interactive coding challenges
Challenges specific to the topic at hand, with tests so you can verify your results.
Build with Botie
Botie helps you follow the pieces of a method before you build a solution yourself. In his workshop, you choose components, connect their inputs, and inspect the results. He also explains quiz answers and offers hints if you get stuck on a programming challenge.
Try the workshop tutorial →