John Hopfield
active c. 1933, Chicago, Illinois, USA
American physicist-biologist who proposed the 'Hopfield network' — an associative-memory model that opened modern machine learning.
Hopfield moved between condensed matter, molecular biology (proving 'kinetic proofreading' in protein synthesis) and computational neuroscience. In 1982 he wrote a recurrent binary-neuron model converging to 'energy minima' — turning neural networks into a statistical-physics problem and reviving a dormant field.
Nobel-winning work
The Hopfield network (1982) assigns each state an Ising-like 'energy': neuron updates always lower it, so memories are local minima the network 'rolls' into — recall by content rather than address. The spin-glass connection allows capacity and error analysis; this energy idea then fed Hinton's Boltzmann machine and inspired deep learning.
Workplaces: Princeton University, Caltech
United States