The Perceptron Convergence Theorem
Day 12 of the ML: Teach by Doing Project is the last part of our 3 part series on the Perceptron.
In this lecture, we learn about:
(a) Linear Separability of a dataset
(b) Margin of a point and margin of a dataset
(c) The Perceptron Convergence Theorem
How cool is it that under certain conditions, the perceptron is guaranteed to converge?
Not just that, we can also predict the maximum number of iterations in which it can converge.
I wish such awesome mathematical behaviour was displayed by modern ML algorithms.
Learn all about margins and the perceptron convergence theorem here:
Here are my lecture notes for this lecture: Link
Stay tuned for Day 13!