Training

I'm attempting to teach myself some Python. As a (particularly useless) exercise, I thought I'd try to write a perceptron neural net. I think I've got it working! As far as I can tell this program works for any linearly separable goal. I'm sort of proud of it, so I thought I'd post the listing here:

# perceptron.py: single-neuron perceptron neural network
# (c) Pericles v. 2.0, 6/12/2008

import sys
import random

input = [[0, 0], [0, 1], [1, 0], [1, 1]]
# randomize on (0, 1] both state vector elements
state = [random.random(), random.random()]
# randomize on (0, 1] bias value
bias = random.random()
# initial output list
output = [0, 0, 0, 0]
# desired output list: simple OR (this program can learn any 
# linearly separable set)
goal = [0, 1, 1, 1]

# hard limit evaluation function: if dot product of the state and input
# vectors is greater than or equal to the bias, then neuron fires
def eval(row, col, bias):
 product = 0
 for i in range(len(row)):
  product += row[i] * col[i]
 if product >= bias:
  fire = 1
 else:
  fire = 0
 return fire

# transition function: compare the binary values of the current output
# element to the goal element, and adjust the state vector if different
def transition(state, input, output, goal, bias):
 b = bias
 s = state
 # for every output element that is not equal to its corresponding goal
 # element, adjust the state vector and the bias
 if output != goal:
  for i in range(len(s)):
   s[i] += (goal - output) * input[i]
  b -= goal - output
 return s, b

# main
print "Simple OR:", goal
print "----------------------------"
print "Output         State    Bias"
print "----------------------------"
for j in range(100):
 # iterate the 4 input vectors through the evaluation function
 for i in range(len(input)):
  output[i] = eval(state, input[i], bias)
 # if the output vector has reached the goal vector, then quit
 if output == goal:
  print output, " ", state, " ", bias
  print "Training complete!"
  sys.exit()
 # iterate the output values through the transition function, and
 # retrieve the adjusted values for the state vector and the bias
 for i in range(len(output)):
  state, bias = transition(state, input[i], output[i], goal[i], bias)
  print output, " ", state, " ", bias