Hopfield Network — Capacity & Spurious States

The Hopfield network stores binary patterns as energy minima via Hebbian learning. The critical storage capacity is α_c ≈ 0.138N patterns (Amit, Gutfreund, Sompolinsky 1985). Beyond capacity, retrieval fails and spurious mixture states appear. Watch the network converge from noisy initial states — can it recover the stored memory?

N: 100 Patterns p: 5 Noise: 0.20

Build a network to begin.