Neural networks huma sistemi tal-komputazzjoni inspirati ħażin mill-moħħ — komposti minn nodes (neurons) interkonness organizzati f'saffi li jitgħallmu patterns mid-data. Huma l-fondazzjoni tal-deep learning u tal-AI modern.
X'inhu neural network
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)
