Module: Neuronet::Tao

Defined in:
lib/neuronet.rb

Overview

A Perceptron Hybrid, Tao directly connects the output layer to the input layer.

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.bless(myself) ⇒ Object

Tao.bless connects the network’s output layer to the input layer, extends it with Tao, and modifies the learning constant if needed.



472
473
474
475
476
477
478
479
480
481
 
# File 'lib/neuronet.rb', line 472

def self.bless(myself)
  # @out directly connects to @in
  myself.out.connect(myself.in)
  myself.extend Tao
  # Save current learning and set it to muk(1).
  l, m = myself.learning, myself.muk
  # If learning was lower b/4, revert.
  myself.learning = l if l<m
  return myself
end

Instance Method Details

#mu ⇒ Object

Tao’s extra connections adds to mu.



465
466
467
468
469
 
# File 'lib/neuronet.rb', line 465

def mu
  sum = super
  sum += self.first.length * self.last.length
  return sum
end