Optimization for Machine Learning
Fitting a model means minimizing something. The objective’s geometry decides everything else.
-
-
Statistical Inference for Machine Learning
Machine learning borrows its estimators from statistics and too often leaves the uncertainty behind.
-
Calculus
Training a network is following a slope downhill. This is the calculus that defines the slope.
-
Probability for Machine Learning
Most modeling mistakes are probability mistakes wearing a modeling costume.
-
Linear Algebra for Machine Learning
If you have looked at a spreadsheet, you have looked at the central object of machine learning. Everything else is built on that.
-
Python and NumPy Basics
Every code block in this series is NumPy or PyTorch. This is the array vocabulary they both assume.
-
Machine Learning Fundamentals
A function fitted to a sample fits that sample better than it fits anything else. Everything difficult follows from that.