AI and ML
BE-CSE-VI
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 1
Perceptron
• The perceptron is one of the earliest and simplest forms of neural networks,
proposed by Frank Rosenblatt in 1957.
• It's a binary classifier used for supervised learning of binary classification
tasks. The perceptron takes multiple input values, processes them linearly,
and produces a single binary output.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 2
How the perceptron works?
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 3
How the perceptron works?
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 4
Limitations
• One limitation of the perceptron is that it can only learn linearly separable
patterns. If the data is not linearly separable, the perceptron may not
converge to a solution. However, for linearly separable problems, the
perceptron is guaranteed to converge.
• Despite its simplicity and limitations, the perceptron laid the foundation for
more complex neural network architectures and learning algorithms. It
played a significant role in the development of artificial neural networks and
machine learning as a field.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 5
Neural Networks
• Neural networks are a type of machine learning model inspired by the
structure and function of the human brain.
• They are composed of interconnected nodes, called neurons, organized into
layers. Each neuron receives input signals, processes them, and then
transmits an output signal to the next layer of neurons.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 6
Neural Networks
• The basic building block of a neural network is the perceptron, which takes
multiple inputs, applies weights to them, sums them up, and then passes the
result through an activation function to produce an output.
• By adjusting the weights and biases of the perceptrons, neural networks can
learn to approximate complex functions and make predictions from data.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 7
Neural Networks Organization
• Neural networks are typically organized into layers:
1. Input Layer: Receives input data and passes it to the next layer.
2. Hidden Layers: Intermediate layers between the input and output layers. They
perform computations on the input data.
3. Output Layer: Produces the final output of the network.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 8
Neural Networks Organization
• Deep neural networks have multiple hidden layers, allowing them to learn
complex representations of data.
• The process of adjusting the weights and biases of a neural network to
minimize the difference between its predicted outputs and the true outputs is
known as training.
• This is typically done using optimization algorithms like gradient descent and
backpropagation.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 9
Why neural networks are popular?
• Neural networks have gained popularity due to their ability to learn from
large amounts of data and solve a wide range of tasks, including image and
speech recognition, natural language processing, and even playing games like
Go and chess at a superhuman level.
• They are a fundamental component of modern artificial intelligence and
machine learning systems.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 10
One-layer Neural Network
• A one-layer neural network, often called a single-layer perceptron, is the
simplest form of a neural network.
• It consists of only one layer of neurons, which directly connect the input to
the output without any intermediate hidden layers.
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 11
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 12
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 13
Note
• It's important to note that a single-layer perceptron can only learn linearly
separable patterns.
• For more complex patterns, you need to use multiple layers (multilayer
perceptron) or other types of neural networks like convolutional neural
networks (CNNs) or recurrent neural networks (RNNs).
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 14
Thanks
Notes: Dr Mumtaz Ali Kaloi CSE Dptt 15