05 Mar 2024 13:39

A more than usually inadequate palceholder.

One important question: why does gradient descent work so well in machine learning, especially for neural networks?

See also: Calculus of Variations and Optimal Control Theory; Computation; Control Theory; Decision Theory; Economics; Evolutionary Computation; Learning in Games; Learning Theory; Low-Regret Learning; Math I Ought to Learn; Planned Economies; Stochastic Approximation