 # Artificial Neural Network - Effect of momentum on learning

All codes used in this tutorial can be found at my Github repository

Before proceeding with this tutorial, prefer to go through my previous tutorial about Artificial Neural Network.

Code Compatibility : Python 2.7 , tested on ubuntu 16.04

As we saw in previous post, Network took pretty long time (~700 epochs) to converge. This is because it was very raw network without any optimization. In present tutorial we will use momentum while making update to weights. Momentum is known to converge network faster.

Momentum($\beta$) term can be represented as :

where $\beta$ is a positive number (1.01 - 1.02) called the momentum constant.  Typically, the momentum constant is set to 1.01 Above said equation is called the generalized delta rule.

To introduce new momentum term we will change our base code. In present implementation we are $\beta$ = 1.01

import math
"""
defining XOR gate, [x1, x2 , y]
"""
XOR = [[0, 1, 1],  [1, 1, 0],  [1, 0, 1],  [0, 0, 0]]

# initializing weights
w13 = 0.5
w14 = 0.9
w23 = 0.4
w24 = 1.0
w35 = -1.2
w45 = 1.1
t3 = 0.8
t4 = -0.1
t5 = 0.3

# defining learning rate
alpha = 0.5
# initializing squaredError
squaredError = 0
# initializing error per case
error = 0
# defining epochs
Epochs = 2000
count = 0
beta = 1.001
# run this repeatedly for number of Epochs
for j in range(Epochs):
print "squaredError", squaredError
# initializing squaredError per epoch
squaredError = 0
for i in range(4):  # iterating through each case for given iteration
"""
calculating output at each perceptron
"""
y3 = 1 / (1 + math.exp(-((XOR[i] *w13) + (XOR[i] *w23 - t3))))
y4 = 1 / (1 + math.exp(-(XOR[i] *w14 + XOR[i] *w24 - t4)))
y5 = 1 / (1 + math.exp(-(y3 * w35 + y4 * w45 - t5)))
"""
calculating error
"""
error = XOR[i] - y5
"""
calculating partial error and change in weight for output and hidden perceptron
"""
del5 = y5 * (1 - y5) * error
dw35 = alpha * y3 * del5
dw45 = alpha * y4 * del5
dt5 = alpha * (-1) * del5
"""
calculating partial error and change in weight for input and hidden perceptron
"""
del3 = y3 * (1 - y3) * del5 * w35
del4 = y4 * (1 - y4) * del5 * w45
dw13 = alpha * XOR[i] *del3
dw23 = alpha * XOR[i] *del3
dt3 = alpha * (-1) * del3
dw14 = alpha * XOR[i] *del4
dw24 = alpha * XOR[i] *del4
dt4 = alpha * (-1) * del4
"""
calculating weight and bias update
"""
w13 = (beta * w13) + dw13
w14 = (beta * w14) + dw14
w23 = (beta * w23) + dw23
w24 = (beta * w24) + dw24
w35 = (beta * w35) + dw35
w45 = (beta * w45) + dw45
t3 = (beta * t3) + dt3
t4 = (beta * t4) + dt4
t5 = (beta * t5) + dt5
"""
Since y5 will be in float number between (0 - 1)
Here we have used 0.5 as threshold, if output is above 0.5 then class will be 1 else 0
"""
if y5 < 0.5:
class_ = 0
else:
class_ = 1
"""
uncomment below line to see predicted and actual output
"""
# print ("Predicted",class_," actual ",XOR[i])
"""
calculating squared error
"""
squaredError = squaredError + (error * error)
if squaredError < 0.001:
# if error is below   0.001, terminate training (premature termination)
break



Upon plotting squared error obtained with  simple XOR gate  and XOR gate with momentum term, we will get below given graph. Figure 1. Improvement in rate of learning after application of momentum term to learning

Share: