Machine Learning Prerequisites Map

201 concepts organized by true dependency order across CS, mathematics, and machine learning - with learning cost metrics and interactive visualizations.

Why This Exists

Most ML learning paths follow academic tradition - Calculus I, then Calculus II, then Linear Algebra. But gradient descent does not care about course catalogs. It needs partial derivatives, matrix multiplication, and the chain rule simultaneously. This map shows what actually depends on what.

201 concepts organized by true dependency order. Every concept connects to prerequisites below it and unlocks capabilities above it. Find where you are, see what you need next, skip what you do not.

What Makes This Different

Learning Cost Metrics

Each concept includes quantified metrics: how many atomic elements you need to learn, the total prerequisite depth, and how many downstream concepts it unlocks. High fan-out concepts give better ROI than dead ends.

Interactive Visualizations

34 concepts include animated canvas visualizations. These are a work in progress. If something looks rough or you have feedback - let me know.

Where to Start

If you're new to the math behind ML, start with these high-ROI entry points - concepts with low prerequisite cost but high downstream unlock count:

Coordinate Systems

Cartesian coordinates, plotting points, distance formula.

Limits

Value a function approaches as input approaches a point. Foundation of calculus.

Vectors Introduction

Directed quantities with magnitude and direction. Addition and scalar multiplication.

Feedback Welcome

This is an active project - the dependency graph, visualizations, and lessons are all evolving. If you find something wrong, have a suggestion, or want to say what was useful - admin@templeton.host or LinkedIn.