Spiking Neural Network

Spiking Neural Network (SNN) is a class of artificial neural network that models neuronal communication through discrete spike events rather than continuous firing rates. Unlike conventional artificial neural networks, which propagate real-valued activations through layers, SNNs operate in the temporal domain: each neuron accumulates input until a threshold is reached, at which point it emits a spike and resets. This makes SNNs fundamentally closer to biological neural dynamics, but also harder to train using standard backpropagation methods.

The appeal of spiking networks lies in their energy efficiency: a neuron that fires only when necessary consumes far less energy than one that continuously computes. This has made SNNs a focus of neuromorphic engineering, the design of hardware that mimics biological neural architecture. Whether the added biological fidelity translates to computational advantages remains contested. SNNs may represent a genuine shift in how we think about neural computation — from rate-coded vectors to temporal codes — or they may be a detour into biological detail that engineering does not require.