Cutting Edge DSS Technology Ch-10

Cutting Edge DSS
Technology
Neural Computing &
Intelligent Agents
Objectives
 Neural computing is an approach
that attempts to mimic the manner
in which our brain works. It is the
one of the several approaches of
machine learning. This class is
dedicated to neural computing
covering the following topics
Neural Computing: The Basics
*Artificial Neural Networks (ANN)

*Mimics How Our Brain Works
*Machine Learning
Machine Learning: An
Overview
 ANN to automate complex decision
making
 Neural networks learn from past
experience and improve their
performance levels
 Machine learning: methods that
teach machines to solve problems
or to support problem solving, by
applying historical cases
Complications
 Many models of learning
 Match the learning model with problem
type
What is Learning?
 Through analogy, discovery, and special
procedures; by observing; or by analyzing
examples
 Is a support area of AI
Learning as Related to AI
 Learning systems demonstrate
interesting learning behaviors
 Learningin AI involves the

manipulation of symbols (not
numeric information)
Machine Learning Method
Examples
 Neural Computing
 Inductive Learning
 Case-based Reasoning and Analogical Reasoning
 Genetic Algorithms
 Statistical Methods
 Explanation-based Learning
Neural Computing
 Computers that mimic certain processing
capabilities of the human brain
 Knowledge representations based on
 Massive parallel processing
 Fast retrieval of large amounts of information
 The ability to recognize patterns based on historical

cases Neural Computing = Artificial Neural
Networks (ANNs)
The Biology Analogy
Biological Neural Networks
Neurons: brain cells
Nucleus (at the center)
Dendrites provide inputs
Axons send outputs
Synapses increase or decrease

connection strength and cause
excitation or inhibition of subsequent
neurons
Artificial Neural Networks
(ANN)
 A model that emulates a biological neural
network
 Software simulations of the massively parallel

processes that involve processing elements
interconnected in a network architecture
 Originally proposed as a model of the human

brain’s activities
 The human brain is much more complex

Three Interconnected Artificial
Neurons
Neural Network Fundamentals
 Components and Structure
Processing Elements
 Network
 Structure of the Network
 Processing Information by the Network

Inputs
 Outputs
 Weights
 Summation Function
Learning: Three Tasks
 Compute Outputs
 Compare Outputs with Desired

Targets
 Adjust
Weights and Repeat the
Process
Neural Network Fundamentals
 Set the weights by either rules or randomly
 Set Delta = Error = actual output minus desired

output for a given set of inputs
 Objective is to Minimize the Delta (Error)
 Change the weights to reduce the Delta
 Information processing: pattern recognition

Different learning algorithms
Neural Network Application
Development
 Preliminary steps of system development are
done
 ANN Application Development Process

1. Collect Data
2. Separate into Training and Test Sets
3. Define a Network Structure
4. Select a Learning Algorithm
5. Set Parameters, Values, Initialize Weights
6. Transform Data to Network Inputs
7. Start Training, and Determine and Revise
Weights
8. Stop and Test
9. Implementation: Use the Network with New
Cases
NN Development Tools
 Braincel (Excel Add-in)
 NeuralWorks
 Brainmaker
 PathFinder
 Trajan Neural Network Simulator
 NeuroShell Easy
 SPSS Neural Connector
 NeuroWare
Neural Network Hardware
 Massive parallel processing greatly
enhances performance Faster general
purpose computers
 General purpose parallel processors
 Neural chips
 Acceleration boards
Benefits of Neural
Networks
 Pattern recognition, learning, classification, generalization and abstraction, and
interpretation of incomplete and noisy inputs
 Character, speech and visual recognition
 Can provide some human problem-solving characteristics
 Can tackle new kinds of problems
 Robust
 Fast
 Flexible and easy to maintain
 Powerful hybrid systems

Limitations of Neural
Networks
 Do not do well at tasks that are not done well
by people
 Lack explanation capabilities
 Limitations and expense of hardware

technology restrict most applications to
software simulations
 Training time can be excessive and tedious

Usually requires large amounts of training and
test data
Neural Computing Use
 Neural Networks in Knowledge Acquisition Fast
identification of implicit knowledge by
automatically analyzing cases of historical data
 ANN identifies patterns and relationships that

may lead to rules for expert systems
 A trained neural network can rapidly process

information to produce associated facts and
consequences
Neural Networks For Decision
Support
 Inductive means for gathering, storing, and using
experiential knowledge
 Neural network-based DSS to appraise real estate in New

York (90% accurate)
 Forecasting
 ANN in decision support: Easy sensitivity analysis and partial

analysis of input factors
 ANN can expand the boundaries of DSS

Summary
 Machine learning
 The human brain is composed of billions of cells

called neurons
 Artificial neural networks learn from historical

cases
 Parallel processing can improve the training

and running of neural networks
 Neural network has got potential to enhance

DSS

Cutting Edge DSS Technology Ch-10

Uploaded by

Copyright:

Available Formats

Cutting Edge DSS Technology Ch-10

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cutting Edge DSS Technology Ch-10

Uploaded by

Copyright:

Available Formats

Cutting Edge DSS

*Artificial Neural Networks (ANN)

 Learningin AI involves the

 Case-based Reasoning and Analogical Reasoning

 Knowledge representations based on

 Massive parallel processing

 Fast retrieval of large amounts of information

 The ability to recognize patterns based on historical

Nucleus (at the center)

Dendrites provide inputs

Axons send outputs

Synapses increase or decrease

 Software simulations of the massively parallel

 Originally proposed as a model of the human

 The human brain is much more complex

 Structure of the Network

 Processing Information by the Network

 Compare Outputs with Desired

 Set Delta = Error = actual output minus desired

 Objective is to Minimize the Delta (Error)

 Change the weights to reduce the Delta

 Information processing: pattern recognition

 ANN Application Development Process

 Trajan Neural Network Simulator

 SPSS Neural Connector

 General purpose parallel processors

 Character, speech and visual recognition

 Can provide some human problem-solving characteristics

 Can tackle new kinds of problems

 Flexible and easy to maintain

 Powerful hybrid systems

 Lack explanation capabilities

 Limitations and expense of hardware

 Training time can be excessive and tedious

 ANN identifies patterns and relationships that

 A trained neural network can rapidly process

 Neural network-based DSS to appraise real estate in New

 ANN in decision support: Easy sensitivity analysis and partial

 ANN can expand the boundaries of DSS

 The human brain is composed of billions of cells

 Artificial neural networks learn from historical

 Parallel processing can improve the training

 Neural network has got potential to enhance

You might also like