Fundamentals of
Neural Network
Dr. Sheetal Mapare
Model Of A Biological
Neuron
Dendrite: Receives signals from
other neurons
Soma: Processes the information
Axon: Transmits the output of
this neuron
Synapse: Point of connection to
other neurons
Model Of A Biological Neuron
McCulloch-Pitts Neuron
It may be divided into 2 parts.
• The first part, g takes an
input performs an
aggregation and
• based on the aggregated
value the second
part, f makes a decision.
Threshold (θ theta):
•A fixed value that the weighted sum must exceed for the
neuron to "fire" or activate
McCulloch-Pitts Neuron
• Input: x1=1,x2=1
• Weights: w1=0.7 (excitatory), w2=−0.5 (inhibitory)
• Threshold (θ): 0.2
• The net input is:
• Net Input =(x1⋅w1)+(x2⋅w2)=(1⋅0.7)+(1⋅−0.5)=0.7
• If the threshold is 0.2, the neuron just meets the threshold and
"fires."
McCulloch-Pitts Neuron
The model is inspired by biological Artificial Neurons (McCulloch-Pitts or Modern Models)
neurons: In artificial neural networks, excitatory and inhibitory
signals are modeled using weights:
1.Inputs: Represent dendrites
1.Excitatory Weights:
receiving signals.
1. Positive weights (w>0) amplify the input signal,
2.Weights: Correspond to the increasing the likelihood of neuron activation.
strength of synaptic 2. Example: If an input is 1 and the weight is +0.5,
connections.
the contribution to the summation is +0.5, which
3.Summation: Mimics how a adds positively to the net input.
neuron integrates inputs. 2.Inhibitory Weights:
4.Threshold: Reflects the firing 1. Negative weights (w<0) suppress the input signal,
threshold of a neuron. reducing the likelihood of neuron activation.
5.Output: Indicates whether the 2. Example: If an input is 1 and the weight is −0.5,
neuron fires (sends a signal via the contribution to the summation is −0.5, which
its axon). subtracts from the net input.
Boolean Functions Using M-P Neuron
This representation just
denotes that, for the
boolean
inputs x_1, x_2 and x_
3 if
the g(x) i.e., sum ≥ th
eta, the neuron will fire
otherwise, it won’t.
NOT and NOR Gate
Nand Gate
Perceptron
• Definition of Perceptron
• The Perceptron is a type of artificial neuron introduced by Frank Rosenblatt in 1958. It is a simple supervised learning algorithm for binary classification
tasks. A perceptron takes multiple input values, computes their weighted sum, applies an activation function (typically a step function), and produces an
output.
Working of a perceptron
Example of AND Gate
Learning with Weight Adjustments
Learning with Weight Adjustments
Predicted output (y): 1 - Target output (t): 1
Learning with Weight Adjustments
Solve
the target output y=0
Solution
