Chapter 40 — Further Reading: Where to Go Next

This is the last "further reading" of the book, so it is also a whole-book reading list: the advanced texts that continue each road §40.4 named, organized by destination. Pick by where you want to go. Annotations say what each source is good for and which chapters of this book prepare you for it.

Advanced linear algebra — finish the core, then deepen it

  • Gilbert Strang, Introduction to Linear Algebra (and Linear Algebra and Learning from Data). The application-first, geometry-forward tradition this book grew out of. Learning from Data in particular is the natural next step for the data-science reader: it carries the SVD (Chapter 30), PCA (Chapter 32), and least squares (Chapter 17) directly into modern machine learning, including a careful treatment of the randomized and low-rank methods of §40.4. Strang's MIT OpenCourseWare lectures (18.06 and 18.065) are free and superb.
  • Sheldon Axler, Linear Algebra Done Right. The abstract, proof-led complement — vector spaces and linear maps first, determinants last, the spectral theorem developed with unusual care. This is the book for the math major who loved Chapters 5, 34, and 35 and wants the theory tightened. It is the cleanest bridge from this book into the abstract roads (multilinear algebra, functional analysis).
  • Carl Meyer, Matrix Analysis and Applied Linear Algebra. A comprehensive reference that sits between Strang and Axler, with thorough treatments of the four fundamental subspaces (Chapters 13–14) and excellent worked applications.

Numerical & randomized linear algebra — computing at scale

  • Lloyd N. Trefethen & David Bau III, Numerical Linear Algebra. The classic, and the right next book after Chapter 38. Built around the SVD and QR, conditioning and stability, written in short, beautiful "lectures." It will make the randomized-SVD sketch of §40.4 rigorous and teach you the algorithms that actually run inside numpy and LAPACK.
  • Gene Golub & Charles Van Loan, Matrix Computations. The encyclopedic reference for serious numerical work — the book to own if you implement matrix algorithms professionally.
  • Halko, Martinsson & Tropp, "Finding Structure with Randomness" (survey, 2011). The foundational survey of randomized methods for low-rank approximation — the theory behind the §40.4 randomized-SVD snippet and Exercise 40.21. [verify: title/year.]

Tensors & multilinear algebra — the deep-learning road

  • Kolda & Bader, "Tensor Decompositions and Applications" (survey, 2009). The standard entry point to tensor decompositions (CP, Tucker) referenced in §40.2 and Case Study 40.1, honest about where the matrix-SVD intuition does and does not transfer. [verify: title/year.]
  • Goodfellow, Bengio & Courville, Deep Learning. Its early chapters give the cleanest textbook account of why deep learning is built on tensors and contractions (§40.2); the linear-algebra and probability primers map directly onto Chapters 2–18 of this book. Freely available online.
  • The PyTorch / TensorFlow documentation, especially einsum. Not a textbook, but the place to practice the contractions of §40.2 and Case Study 40.1. Reading these docs with this chapter in hand, you will recognize every operation.

Functional analysis & infinite dimensions — the physics road

  • Kreyszig, Introductory Functional Analysis with Applications. The gentlest serious introduction — metric spaces, Banach and Hilbert spaces, bounded operators (§40.3) — assuming little beyond the linear algebra and basic analysis you now have. The right book to make Chapters 22 and 34 rigorous.
  • Reed & Simon, Methods of Modern Mathematical Physics, Vol. I: Functional Analysis. The standard graduate reference for the operator theory underlying quantum mechanics; demanding, and the destination rather than the on-ramp.
  • The DataField quantum-mechanics book, particularly the infinite-dimensional spaces in quantum mechanics treatment — the applied, physics-first counterpart that shows why the functional analysis matters, continuing the qubit anchor of Case Study 40.2.

Quantum computing — the qubit's road onward

  • Nielsen & Chuang, Quantum Computation and Quantum Information. The field's standard text ("Mike & Ike"). Its opening chapters are, essentially, the linear algebra of Case Study 40.2 done carefully — qubits as unit vectors, gates as unitaries, the tensor product of state spaces — before moving to algorithms (Shor, Grover) and error correction. You are well prepared for its first chapters today.
  • John Preskill's lecture notes (Caltech Ph219) and Qiskit's free textbook — excellent free, hands-on resources; Qiskit in particular lets you run the Hadamard, CNOT, and Bell-state code of Case Study 40.2 on a simulator. [verify: course numbers/availability.]

Optimization — the machine-learning & operations road

  • Stephen Boyd & Lieven Vandenberghe, Convex Optimization. The standard, and the natural continuation of Chapters 17–19 and 28. Built on positive-definiteness, quadratic forms, and projection — the linear algebra of §40.4's optimization paragraph — and freely available online, with Boyd's lectures.
  • Boyd & Vandenberghe, Introduction to Applied Linear Algebra (VMLS). A gentler, application-saturated linear-algebra-for-data book that doubles as a bridge into optimization; free online.

The other DataField roads

  • Multivariable calculus — the other half of the toolkit: the derivative as a linear map (the Jacobian), the gradient and Hessian behind §40.4's optimization. Pair this with any of the above.
  • The DataField data-science sequence (intro → intermediate → advanced) and the statistics book — where PCA, regression, recommenders, and covariance (Chapters 17, 30, 32, 33) become full methods on real data, and where the probabilistic half of the story is supplied.
  • The discrete-mathematics book — graphs, which become matrices the moment you compute with them (the spectral graph theory of §40.4).

How to actually use this list

Do not read it top to bottom. Pick the one road that matches your destination (the §40.4 FAQ helps), choose its entry-level text and pair it with Strang or Axler, and — most importantly — build something with the toolkit you wrote while you read (the §40.7 Build Your Toolkit callout and Exercises 40.23–40.25 point the way). Fluency in linear algebra does not come from re-reading theorems; it comes from using the subject on a problem you care about, which is exactly where this book has been pointing you all along.