Mathematics for Machine Learning
For developers and aspiring ML engineers who want the maths behind machine learning explained with worked numbers and runnable NumPy code. You will be able to read the linear algebra, calculus and statistics inside ML libraries and papers, and use them to debug, compare and tune real models.
Course Content
Read data as vectors and matrices, solve linear systems without numerical traps, and use eigenvectors to find the directions that matter in a dataset.
Compute derivatives and gradients, push them backwards through a network with the chain rule, and use them to train a model by gradient descent.
Describe uncertainty with random variables and named distributions, and test whether a difference between two models is real or just noise.
Put the earlier maths to work on real models: choosing a loss, deriving linear regression from scratch, and regularising a model so it stops fitting noise.
Build a gradient descent visualiser in Python and watch learning rate, momentum and curvature decide whether a run converges, crawls or blows up.