Skip to content
Ben Kennedy Explainers · Inside a Spiking Neural Network
Events, not streams

Silence is cheap.
Spikes carry the news.

Watch a network that only speaks when something happens. Events move, quiet connections rest, and timing itself carries information. Tune the neuron and see the model respond.

01 / 07 Stage
What’s Happening Now · Reading events over time event driven
Animated diagram of a spiking neural network processing timed events. Events over time
Key idea

Spiking neurons communicate with discrete events whose timing can carry information.

Input events44 / window
Memory half-life7.0 steps
Output spikes3 / window
These settings produce repeated threshold crossings.

01 · Goal

Use timing as part of the message.

A spiking network stays quiet until an event matters. When enough evidence arrives, a neuron emits a brief spike.

Watch Only small events move through the network. Silence carries no update.
Open the technical explanation

Nothing moves
until it matters.

A spiking network spends energy only on events. Evidence accumulates, a threshold decides, and a brief spike carries the result forward. The quiet in between is not waste. It is the design.

Ben Kennedy builds privacy-first AI for regulated industries. Founder of Kennedy Applied Sciences, creator of StratoSort, PhD candidate in Artificial Intelligence.

This explainer connects to my ongoing STAC research on neuromorphic spiking neural networks.