Unit I
Basics- Biological Neuron, Idea of computational
units, McCulloch–Pitts unit and Thresholding
logic, Linear Perceptron, Perceptron Learning
Algorithm, Linear separability, Convergence
theorem for Perceptron Learning Algorithm.
Biological Neurons:
The human brain consists of a large number, more than a billion of
neural cells that process information. Each cell works like a simple
processor. The massive interaction between all cells and their parallel
processing only makes the brain’s abilities possible. Figure 1 represents
a human biological nervous unit. Various parts of biological neural
network(BNN) is marked in Figure 1.
Dendrite: receives signals from other neurons
Synapse: point of connection to other neurons
Soma: processes the information axon: transmits the output of this
neuron
Function:
Receive signals from other neurons.
Transmit signals to other neurons.
Integrate signals from multiple neurons.
Fire an action potential if the signal is strong enough.
Release neurotransmitters that signal other neurons of the nervous
system. They transmit signals from one part of the body to another,
allowing us to sense the world around us, control our movements,
and think.
Signal Transmission: When a biological neuron receives a signal from
another neuron, it opens ion channels in its membrane. This allows ions
to flow into the cell, which changes the electrical potential of the cell. If
the electrical potential of the cell reaches a certain threshold, the
neuron will fire, sending a signal to other neurons.
The most fundamental unit of a deep neural network is called an
artificial neuron
Why is it called a neuron ? Where does the inspiration come from ?
The inspiration comes from biology (more specifically, from the
brain)
biological neurons = neural cells = neural processing units
Idea of Computational Units
Computational units:
Devices that can perform basic mathematical operations.
Can be used to simulate the behavior of biological neurons.
Can be used to build artificial neural networks.
Can be used to solve complex problems.
Can be used to create intelligent machines
Computational units:
A computational unit is a device that can perform basic mathematical
operations, such as addition, subtraction, multiplication, and division.
Artificial neurons:
Artificial neurons are inspired by biological neurons. They are
computational units that can be used to simulate the behavior of
biological neurons.The McCulloch-Pitts unit is a simple model of an
artificial neuron. It has two inputs and one output. The output of the unit
is determined by a threshold function.
Thresholding logic: Thresholding logic is a type of logic that uses a
threshold function to determine the output of a circuit.
Sure, here are five points for each side heading: Type of logic:
o Uses a threshold function to determine the output of a circuit.
o Can be used to implement simple logical operations, such as AND, OR,
and NOT.
o Can be used to build more complex logic circuits.
o Can be used to implement artificial neural networks.
McCulloch Pitts Neuron
The McCulloch-Pitts unit is a simple model of an artificial neuron. It has
two inputs and one output. The output of the unit is determined by a
threshold function
Two inputs:
o Can be either 0 or 1.
o Represent the signals received from other neurons.
One output:
o Can be either 0 or 1.
o Represents the output of the unit
Let us implement some boolean functions using this McCulloch Pitts (MP)
neuron
A single McCulloch Pitts Neuron can be used to represent boolean functions which are
linearly separable
Linear separability (for boolean functions) : There exists a line (plane)
such that all inputs which produce a 1 lie on one side of the line (plane) and all inputs
which produce a 0 lie on other side of the line (plane)
Perceptron
What kind of functions can be implemented using the perceptron? Any
difference from McCulloch Pitts neurons?
Perceptron Learning Algorithm
Overview
1. The perceptron learning algorithm is a simple algorithm for training a
linear perceptron.
2. The algorithm works by iteratively adjusting the weights and
threshold of the perceptron until it can correctly classify all of the
training data.
step1. The algorithm starts with random weights and threshold.
step2. It then presents each input pattern to the perceptron.
step3. If the perceptron misclassifies the input pattern, the weights and
threshold are adjusted.
step4. This process is repeated until the perceptron can correctly
classify all of the training data.
step 5. The algorithm is guaranteed to converge to a solution if the
training data is linearly separable
Proof of Convergence for the Perceptron or the Learning Algorithm
Linearly Separable Boolean Functions
Linear Separability
Definition
1.A set of data points is said to be linearly separable if there exists a
hyperplane that can perfectly classify the data points.
2. A hyperplane is a flat surface that divides the space into two regions.
3. The data points on one side of the hyperplane are classified as one
class, and the data points on the other side of the hyperplane are
classified as the other class.
4. If the data points are linearly separable, then there exists a
hyperplane that can perfectly classify the data points with no
misclassifications.
5. The perceptron learning algorithm can only be used to train a
perceptron to classify data points that are linearly separable.
6. If the data points are not linearly separable, then the perceptron
learning algorithm will not be able to converge to a solution
