Scribd
Upload
4 views

Soft Computing, Hard Computing, Basics of ANN

Uploaded by

anshikam220
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
4 views

Soft Computing, Hard Computing, Basics of ANN

Uploaded by

anshikam220
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 26

APPLICATION

OF SOFT
COMPUTING
Computing is the process of using computer technology to complete a given
goal-oriented task.

COMPUTING  Function(F) : Formal Method /Algorithms/ Mapping Function


Precise solution : exactly or sharply
defined or stated.
FEATURES OF Unambiguous & Accurate : Should be
COMPUTING
only one meaning of any Control Action.

Mathematical Model
HARD COMPUTING

 Hard computing is a traditional computing.

 It requires a precisely stated analytical model and usually a lot of computation time.

 It strictly follows known steps to solve a task as opposed to soft computing which
is heuristic.
 Hard computing is achieved using sequential programs that use binary logic.

 It is deterministic in nature.

 The input data should be exact, and the output will be precise and verifiable.
Precise solution : exactly or sharply defined or stated.

Unambiguous & Accurate : Should be only one meaning of


any Control Action.
FEATURES OF
HARD COMPUTING Mathematical Model
& EXAMPLE
Formally Defined

Example: Numerical Problem, Sorting and searching


Algorithems.
CONTINUE...

Advantages
 Accurate solutions can be obtained

 Faster

Disadvantages
 Not suitable for real world problems

 Cannot handle imprecision and partial truth


WHAT IS SOFT COMPUTING?
 Soft computing is the use of approximate calculations to provide imprecise but
usable solutions to complex real world problems.
 The approach enables solutions for problems that may be either unsolvable or
just too time-consuming to solve with current hardware.
 Soft computing is sometimes referred to as computational intelligence.

 As a field of mathematical and computer study, soft computing has been around
since the 1990s.
USE CASE
 Consumer appliance like AC, Refrigerator, Heaters, Washing machine.

 Food preparation devices are Microwave and Rice cookers.

 Data compression/Image Processing


FEATURE OF SOFT COMPUTING

 Imprecision

 Dynamic

 Uncertainty

 Low Solution Cost

 Do Not required Mathematical Model


TOOLS/COMPONENTS
Some of its principal tools/components includes:
 Neural Network(NN)

 Fuzzy Logic(FL)

 Genetic Algorithm(GA)

These methodologies form the core of soft computing.


GOALS OF SOFT COMPUTING

 The main goal of soft computing is to develop intelligent machines to provide


solutions to real world problems, which are not modeled, or too difficult to model
mathematically
 Delas with Approximation, Uncertainty, Imprecision, and Partial Truth in order to
achieve close resemblance with human like decision making.
HARD COMPUTING VS SOFT COMPUTING
Hard computing Soft computing

 Precision  Imprecision

 Based on binary logic and numerical  Based on fuzzy logic, neural network.
analysis
 Soft computing produces approximate
 Hard computing produces precise results. results.
 Input exact data.  Deals with ambiguous and noisy data.

 Hard computing uses two-valued logic.  Soft computing will use multivalued logic.

 Hard computing performs sequential  Soft computing can perform parallel


computations. computations.
NEURON
BASIC WORKING UNIT OF BRAIN
NEURONS ARE INFORMATION
MESSENGERS. THEY USE
ELECTRICAL IMPULSES AND
CHEMICAL SIGNALS TO TRANSMIT
INFORMATION BETWEEN
DIFFERENT AREAS OF THE BRAIN,
AND BETWEEN THE BRAIN AND
THE REST OF THE NERVOUS
SYSTEM
CONTINUE...
Types of Biological Neuron
 A nerve cell neuron is a special
biological cell that processes Sensory Neurons: taste, smell, hear, see, feel.
information. According to an
Motor Neurons: Movements of muscles or organs.
estimation, there are huge number of
neurons, approximately 1011 with Interneurons: Intermediate neurons found in spinal cord
numerous interconnections, and brain. They pass signals from sensory neurons to
approximately 1015. motor neurons
BRIEF HISTORY OF NEURAL NETWORKS
 Model of a neuron

ARTIFICIAL
NEURAL
NETWORK

Artificial neurons (also called Perceptrons, Units or


Nodes) are the simplest elements or building blocks in a
neural network. They are inspired by biological neurons
that are found in the human brain.
ANN VERSUS BNN

 Before taking a look at the


differences between Artificial
Neural Network ANNANN and
Biological Neural
Network BNNBNN, let us take a
look at the similarities based on
the terminology between these
two.
CONTINUE...
 Dendrites (input) − They are tree-like branches, responsible for receiving the information from other
neurons it is connected to. In other sense, we can say that they are like the ears of neuron.
 Soma (Processing Unit) − It is the cell body of the neuron and is responsible for processing of
information, they have received from dendrites. It sums up all the input signals. It also consist of
threshold unit
 Synapses (weighted connection) - When the sum reaches a threshold value, neuron fires and
generate an output signal. It is point of interconnection of one neuron with other neuron. The amount
of signal transmitted depends upon the strength(synaptic weights) of the connection. Connections
can be inhibitory(decreasing strength) or excitatory(increasing strength) in nature.
 Axon(Output) − It is just like a cable through which neurons send the information which travel down
to other neuron.
COMPARISON OF ANN AND BNN
ARTIFICIAL NEURON AND ITS MODEL
Artificial Neuron: An artificial neuron is mathematical function based on model of biological neurons.

Each artificial neurons has the following main functions


1. Receives input from number of input signals
2. Weight each input separately and sum them up
3. Pass this sum through an activation function to produce output
NETWORK ARCHITECTURE CONTD..
COMPONENTS OF ARTIFICIAL NEURAL NETWORK

Components of NN:

• Input later/Input Values


• Hidden Layer
• Synapse/weighted connection link
• Summing function
• Activation function
• Output layer
COMPONENTS CONT…..
1. Input Values/One input layer:
 Takes raw input from the domin
 No computation is performed at this layer.
 Nodes here just pass on information to the hidden layer.
 It is similar to dendrite in biological neuron

2. Hidden Layer
 Nodes of this layer are not exposed.
 They provide abstraction to the neural network.
 It performs all kind of computation on the features entered through input layer and transfer the result to output layer.

3. Synapse/weighted connection link:


 Each connection link is categorized by a weight or strength(wi).
 Each input xi is multiplied by its weight wi to get synaptic strength.

4. Summing Function:
 The summing function is used to take the sum of all weighted connection link.
 A bias value is also added to this weighted sum.
COMPONENTS CONT…..
5. Activation Function:
 It decides whether or not a neuron is fired based on output value produced.
6. Output/Output Layer:
 It gives the final output/output layer of a neuron/network which can be passed to other neuron in the
network
CALCULATION

Yin= x1w1+x2w2
ACTIVATION FUNCTION
 An Activation Function decides whether a neuron should be activated or not.

 It is a non-linear functions.

 An activation function is just a simple function that changes its inputs into
outputs with a defined range.
 Activation Function also called as transfer functions are equations that define
how the weighted sum of the input of a neural node is transformed into an
output.

You might also like