Machine Learning
Machine learning from its exact arithmetic core: small rational data, validated deterministic diagrams, and clear boundaries around what is computed versus what is not claimed.
20 books, 124 pages, 20 chapters.
Books
- Linear Regression (Least Squares) - 9 pages, 1 chapter
- Ridge Regression, Exactly - 5 pages, 1 chapter
- Logistic Scoring - 6 pages, 1 chapter
- Decision Trees (Gini) - 6 pages, 1 chapter
- k-NN and Boundaries - 6 pages, 1 chapter
- The Perceptron - 6 pages, 1 chapter
- Max-Margin (SVM) - 7 pages, 1 chapter
- Naive Bayes from Counts - 6 pages, 1 chapter
- k-Means, One Honest Step - 6 pages, 1 chapter
- A Neural Net You Can Compute by Hand - 6 pages, 1 chapter
- Backpropagation, Exactly - 6 pages, 1 chapter
- Gradient Descent, One Step - 6 pages, 1 chapter
- Tokenization (BPE) - 6 pages, 1 chapter
- Embeddings - 6 pages, 1 chapter
- Attention, Structurally - 8 pages, 1 chapter
- The Transformer Block - 6 pages, 1 chapter
- Decoding - 6 pages, 1 chapter
- A Transformer, End to End - 6 pages, 1 chapter
- Running a Real Model - 6 pages, 1 chapter
- Evaluation, Honestly - 5 pages, 1 chapter