Backpropagation, step by step

Goal: teach the network to output 0 when it sees the input [1, 1].

The network
  connection (label = weight) less blame more blame b  bias
Round
Output
Target0
Error