Matrix Calculus for Machine Learning and Beyond

1-0-2 units · Undergraduate

Covers a coherent approach to matrix calculus, showing techniques that allow the student to think of a matrix holistically, rather than as an array of scalars; generalize and compute derivatives of important matrix factorizations and other complicated-looking operations; and understand how differentiation formulas must be reimagined in large-scale computing. Discusses "adjoint" or "reverse-mode" differentiation, custom vector-Jacobian products, and how modern automatic differentiation is more computer science than calculus.