Skip to content
kwharrison13.comEssays, books and notes by Kyle Harrison

Algebra, Topology, Differential Calculus, and Optimization Theory for Computer Science and Machine Learning

From kwharrison13.com · Library of Alexandria
Algebra, Topology, Differential Calculus, and Optimization Theory for Computer Science and Machine Learning
AuthorJean Gallier and Jocelyn Quaintance
FormatBook
OriginalRead the original

TL;DR — Gallier and Quaintance’s free 2,200-page Penn textbook on the mathematics behind machine learning, running from groups and linear algebra to optimization and support vector machines.

The book is organized in ten parts across 56 chapters. Part I, Linear Algebra, takes up roughly the first 750 pages: vector spaces, determinants, LU and Cholesky factorization, norms, iterative solvers, Euclidean and Hermitian spaces, eigenvalues, spectral theorems, graph Laplacians and spectral graph drawing, the singular value decomposition and pseudo-inverses, with side chapters on Haar wavelets, unit quaternions and the finite element method. Parts II to IV cover affine and projective geometry, the geometry of bilinear forms, and abstract algebra (polynomial ideals, Noetherian rings, tensor and exterior algebras, modules over a PID, normal forms). Part V covers topology and differential calculus, including a short detour on fractals.

The second half turns to optimization: extrema, Newton’s method, quadratic optimization and Schur complements (Part VI); convex sets, linear programming, the simplex algorithm and duality (Part VII); Hilbert spaces, general optimization theory, nonlinear optimization, subgradients and ADMM (Part VIII). Part IX, Applications to Machine Learning, covers positive definite kernels, soft-margin support vector machines, ridge regression, the lasso and elastic net, and ν-SV regression, followed by appendices, a bibliography and an index. Most chapters end with a summary and problems.

The authors are in the Department of Computer and Information Science at the University of Pennsylvania, and the book is posted free on Gallier’s faculty page. It is a living manuscript: this copy is dated September 29, 2026, and the one-page Introduction chapter is still blank.

What it is: A rigorous, proof-based textbook released by its authors without peer review. It is a reference to consult by chapter rather than read cover to cover, and it assumes mathematical maturity.

Where this came from

2220 pages. A copy is archived locally against link rot; the header links the original source.

Curated by Kyle Harrison. Organized by Claude. Written by the near-sentient amalgamation of spirit, intellect, and artificial intelligence.