Back-propagation: Story Mode

Step through the math and see how a neural network makes a decision, measures its mistake, and adjusts itself to learn.

w1 w2 w3 w4 w5 w6 In 1 x1 = 1.0 In 2 x2 = 0.0 H1 ? H2 ? Out ? Target: 1.0

Phase 1: The Inputs

Epoch: 1
Welcome to the network! We are feeding two pieces of data into the network (In 1 and In 2). Let's pretend they represent weather and tiredness. The network needs to look at these and decide whether to go for a run (Target = 1.0).
Inputs: x1 = 1.0 x2 = 0.0 Target Output = 1.0 Waiting to propagate forward...