Spiking Neural Network
A neural network architecture that simulates the biological neuron's integrate-and-fire mechanism. Neurons emit discrete spikes only when membrane potential exceeds a threshold and remain silent otherwise, yielding sparse computation. This behavior is naturally compatible with event-camera output, making spiking neural networks suitable for low-power edge inference applications.
Timeline (last 90 days)
A compact spiking neural network with 93,000 parameters has an inference energy consumption 15 times lower than the reference solution, at the cost of a 0.073 decrease in macro F1.
Neurons in a spiking neural network emit discrete spikes only when the membrane potential accumulates beyond a threshold, remaining silent otherwise.
Neuromorphic systems typically employ spiking neural networks (SNNs), which convey information using discrete spike events rather than continuous floating-point numbers.
All Facts (3)
A compact spiking neural network with 93,000 parameters has an inference energy consumption 15 times lower than the reference solution, at the cost of a 0.073 decrease in macro F1.
50%UnverifiedNeurons in a spiking neural network emit discrete spikes only when the membrane potential accumulates beyond a threshold, remaining silent otherwise.
50%UnverifiedNeuromorphic systems typically employ spiking neural networks (SNNs), which convey information using discrete spike events rather than continuous floating-point numbers.
50%