Engine preview
A ten-node roadmap that exercises the graph, the path, progress, bookmarks and timestamped notes end to end. The structure is ours; the videos are well-known public explainers rather than picks from the Phase 4 sourcing protocol.
This roadmap is a preview, not a finished one.
It exists so the roadmap engine — the graph, the path, progress, bookmarks and timestamped notes — can be built and reviewed against real content. The structure is ours and the videos are well-known public explainers, but they have not been through the sourcing and review protocol a shipped roadmap goes through. So it sits apart from the catalog, and search engines are told not to index it.
The formula library is finished and live; the catalog says which roadmaps are queued.
The maths underneath
- VectorsWhat a vector is from three directions at once — a list of numbers, an arrow in space, and a thing you can add and scale.
- Dot productsTwo vectors in, one number out — and why that number keeps appearing wherever similarity or projection is being measured.
- Matrix multiplicationMatrices as transformations, and multiplication as doing one transformation after another.
- DerivativesThe derivative as a rate of change you can see, and the chain rule as the reason deep networks are trainable at all.
How a network learns
- What is a neural network?Layers, weights, biases and activations — the whole forward pass, drawn on one handwritten digit.
- Gradient descentThe cost function, the gradient as the direction of steepest ascent, and learning as repeatedly stepping the other way.
- BackpropagationHow a single training example votes on every weight in the network, and why the votes are averaged rather than applied.
- Backpropagation calculusThe same algorithm in symbols — partial derivatives, the chain rule, and where each term in the expression comes from.