Neural Network
A model of interconnected 'neurons' that learns by adjusting connection strengths.
Loosely inspired by the brain, layers of simple units pass signals forward, each weighted connection tuned during training to reduce error. Stacking many layers ('deep learning') lets the network learn very complex patterns in images, audio, and language.
Universal approximators in theory; training via backpropagation and gradient descent, with architecture (layers, activations) and regularization shaping generalization.
MLPs, CNNs for vision, RNNs for sequences, the perceptron, Keras models.
Like a rumor spreading through a crowd where each person amplifies or quiets it: the final message depends on how strongly everyone passes it along, and you 'train' the crowd by correcting them over time.