View source for Spiking Neural Network

'''Spiking Neural Network''' (SNN) is a class of [[Artificial Neural Network|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|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.

[[Category:Technology]]
[[Category:Systems]]