Neurale netværk er computersystemer, der er løst inspireret af hjernens funktioner — bestaende af sammenkoblede knuder (neuroner) organiseret i lag, som lærer mønstre fra data. De er grundlaget for deep learning og moderne AI.
Hvad et neuralt netværk er
NEURAL NETWORK → layers of connected NODES (neurons) that process data:
→ INPUT layer (receives data) → HIDDEN layer(s) (process/transform) → OUTPUT layer (result)
→ each connection has a WEIGHT; neurons combine inputs, apply an activation function
→ the network LEARNS by adjusting weights to map inputs to correct outputs
→ loosely inspired by the brain's neurons (but a mathematical model)
