Skip to content
Free public beta until December 4, 2026: every course is open, no card needed. After that, the first chapter of each course stays free and full access is a one-time $39 ($19 for up to 100 founding members chosen for helpful feedback). Your progress is kept. How the beta works

machine-learning

Data as vectors and matrices

taught

Reading one data point as a feature vector and a whole dataset as a matrix, the representation every ML method in this course assumes.

Taught in

Requires

Required by

Local knowledge map

Local knowledge map

Binary logistic regression Data leakage, spotted during EDA Generalization Modern embeddings and semantic search Multiple linear regression Running PCA in practice Text as data: tokenization, bag-of-words, and… The soft margin and cost parameter The validation-set approach and its weaknesses Time series features: lags and rolling statistics t-SNE and UMAP Feature creation: polynomial, interaction, binning,… Features, samples, and training/validation/test splits PCA: variance maximization and reconstruction error Simple linear regression and least squares The maximal margin classifier Data as vectors and matrices Matrix multiplication Optimization vocabulary for ML: gradient descent and… The batch dimension and broadcasting, as engineering convention Writing a system in matrix form nn.Linear is Wx + b np.dot and @ compute the dot product and the matrix product Matrix-vector multiplication two ways: the row picture…

Click to expand

Solid arrows are required prerequisites; dashed are recommended or helpful. Lower rows are more foundational. Click any box to jump to that concept.