Neural Network

Network Architecture
Decision Boundary (left-click: class 0, right-click: class 1)

Controls

0.15
20

Statistics

Epoch0
Loss-
Accuracy-
Weights-
Loss Over Time
How it works: This feedforward neural network learns via backpropagation with sigmoid activations. Watch the edges change color (blue = negative, gold = positive) and thickness (stronger weight) as the network discovers the decision boundary. The bottom plot shows how the network partitions the 2D input space into two classes.