Scribd
Upload
88 views18 pages

Artificial Neural Network

Uploaded by

Dhruv Bansal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
88 views18 pages

Artificial Neural Network

Uploaded by

Dhruv Bansal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

SOFT COMPUTING

ARTIFICIAL NEURAL
NETWORK

By Dhruv Bansal
INTRODUCTION
An artificial neural network (ANN) is a computing model
inspired by the human brain. It consists of interconnected
nodes (neurons) organized in layers to process data,
recognize patterns, and make decisions. ANNs are widely
used in machine learning for tasks like image recognition,
natural language processing, and predictive analytics.
LAYERS
ANN mainly consist of three layers :
Input layer: The first layer, where data enters network. Each
neuron in this layer represents a feature of the input data.
Hidden Layer: Intermediate layers between the input and output
layers. They process the input data through multiple neurons and
apply transformations. Complex networks can have multiple
hidden layers (deep learning). These layers enable the network to
learn abstract patterns.
Output Layer: The final layer, which produces the network's
result or prediction. The number of neurons in this layer depends
on the type of problem (e.g., one neuron for binary classification,
multiple neurons for multi-class classification).
ARCHITECTURE
TYPES OF TRAINING
METHOD:
There are two fundamental processes in training an
artificial neural network (ANN):
1. Feedforward

2. Backward Propagation
FEEDFORWARD
In the feedforward process, data flows through the network from the
input layer to the output layer:
1. Input Layer: Each feature of the input data is fed to a neuron in the
input layer.
2. Hidden Layers: Each neuron calculates a weighted sum of its inputs,
applies an activation function, and passes the result to the next layer.
3. Output Layer: The final layer produces the network’s prediction or
output.
Feedforward propagation calculates the network’s output based on
current weights but does not involve learning; it’s a single pass through
the network.
BACKWARD PROPAGATION
Backpropagation is the process of adjusting the weights to minimize the
network’s prediction error:
1. Calculate Error: The error or loss is computed by comparing the output
with the actual target value using a loss function.
2. Backpropagate Error: The error is propagated backward through the
network, from the output layer to the hidden layers, using gradients to
measure the sensitivity of each weight to the error.
3. Update Weights: Weights are adjusted using an optimization algorithm
(e.g., gradient descent), aiming to reduce the error. This update process
continues iteratively over many epochs to optimize the network.
Together, feedforward and backpropagation enable an ANN to learn patterns
from data by iteratively refining weights for better accuracy.
PROBLEM
Q. Consider a multilayer feed-forward neural network given below. Let the learning rate be
0.5. Assume initial values of weights and biases as given in the table below. Train the
network for the training tuples (1, 1, 0) and (0, 1, 1), where last number is target output.
Show weight and bias updates by using back-propagation algorithm. Assume that sigmoid
activation function is used in the network.
SOLUTION
Step 1: Initialization
a. Inputs and Outputs for 2 iterations:

b. Initial weight and bias values:

c. Learning rate (l) = 0.5


d. Activation function = Sigmoid function
Step 2: The termination condition for our algorithm, according to the
question, is when we complete 2 iterations.
Iteration 1:
Iteration 2:
Since we completed 2 iterations, according to question, our calculation stops
here.
The updated weight and bias values after second iterations are:
THANK YOU

You might also like