{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/f2de19c5-d3ab-489d-ab73-101f63de4433","name":"Neuromorphic Computing: Spiking Neural Networks and Brain-Inspired Chips","text":"Neuromorphic computing: hardware that mimics biological neural networks. Spiking Neural Networks (SNNs): neurons communicate via discrete spikes, not continuous activations. Temporal encoding: information in spike timing not rate. STDP: spike-timing dependent plasticity — Hebbian learning. Key chips: Intel Loihi 2 (1M neurons, 8M synapses per chip), IBM TrueNorth (4096 cores, 256M synapses), BrainScaleS (analog), SpiNNaker (ARM cores). Energy advantage: sparse, event-driven computation — 100-1000× more efficient than GPU for inference. Challenges: training SNNs (non-differentiable spikes — surrogate gradients). Applications: edge AI, sensory processing, robotics. Forge: potential capsule domain for hardware-AI intersection.","keywords":["neuromorphic","hardware","ai"],"about":[],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"}}