Hopfield Network
A recurrent neural network that serves as associative memory, storing and retrieving patterns.
Overview
A Hopfield network is a form of recurrent neural network introduced by John Hopfield in 1982 that can store and recall patterns as stable states (attractors) of its dynamics. Each neuron is connected to every other neuron with symmetric weights, and the network evolves toward stored patterns when given partial or noisy input.
Key Details
Modern Hopfield networks, updated for the transformer era, can store exponentially many patterns and have connections to the attention mechanism. The 2024 Nobel Prize in Physics was awarded to Hopfield (and Geoffrey Hinton) for foundational work on neural networks, highlighting this architecture's lasting significance.
Related Concepts
recurrent neural network • attention mechanism • boltzmann machine