neural-net-hacking-examples

Roadmap

• Add bias neuron
• Calculate error
• Backprop error
• Train - Neurons update their connect weight values (learning!)
• Simplify training method, use basic stochastic gradient descent
• Add training method or Trainer class
• Standardize Data set shape for training