Scribd
Upload
103 views

Expert System and Neural Networks

Uploaded by

Amit Malhan
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
103 views

Expert System and Neural Networks

Uploaded by

Amit Malhan
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 32

Expert Systems

An expert system is a computer


program that is designed to hold
the accumulated knowledge of
one or more domain experts to
solve problems
Background History
Expert systems first emerged from the
research laboratories of U.S.universities
during the 1960s and 1970s.
They were developed as specialized
problem solvers which emphasized the
use of knowledge rather than data and
general search methods.
Background history
The first expert system to be completed was DENDRAL, developed at stanford university in the late 1960s.

DENDRAL: Used to
identify the structure
of chemical compounds
with consituent elements.
Background history
DENDRAL discovered a number of
structures perviosly unknown to expert
chemists.
As researchers gained more experience
with DENDRAL,led to the development of
Meta-DENDRAL,a learning component
which was able to learn rules from positive
examples, a form of inductive learning.
Background history
PUFF:
Medical system
for diagnosis of
respiratory dieses

PROSPECTOR:
Used by geologists
to identify sites for
drilling or mining
Background history
MYCIN:
Medical system for
diagnosing blood disorders.
First used in 1979

DESIGN ADVISOR:
Gives advice to
designers of
processor chips
Background history
The initial MYCIN’s knowledge base
contained only 200 Rules and vocabulary of
2000 words.
This number was gradually increased to more
than 600 rules by the early 1980s.
This improved MYCIN’s performance leading
to a 65% success record which compared
favorably with experinced physicians with
60% success rate.
Why we use Expert Systems?
Campbell Soup uses large cookers to
cook soups and other can products at
eight plants located throghout the country.
Some cookers hold up to 68,000 cans of
food for short periods of cooking time.
when maintenance problems occur with
the cookers,fault must be found or food
will spoil.
Why we use Expert Systems?
Company had been depending on a single
expert to diagnose.
Expert will retire in a few years taking his
expertise with him.
so, company decided to develop an expert
system.
After some months company developed an
expert system about 150 rules in KB to
diagnose cooker problems.
Why we use Expert Systems?
Capture and preserve irreplaceable
human expertise
Provide expertise needed at a number
of locations at the same time
Provide expertise needed for training of
human experts
Rule based sytem architecture
In this system knowledge encoded in the
form of production rules,i.e.if ….then
rules.e.g.
Inference in production systems is done
by process of chaining either 1.
forwad
2.backward
Nonproduction system
architecture
Instead of rules,thes systems employ more
structured representation schemes like
1.Neural networks 4.Desicion tree
2.Blackboard system
3.Analogical reasoning
4.frame
5.semantic network
Blackboard system
architecture
One of the first blackboard system was
HEARSAY used for speech
understanding system.
More recently ,systems have been
developed to analyze complex scene
and model human cognitive processes.
Blackboard system
architecture

CHINESE ROOM PROBLEM(CONGNITION)


Neural network architecture
Neural network: information
processing inspired by biological
nervous systems, such as our brain.
Structure: large number of highly
interconnected processing elements
working together
Control and Coordination
Reflex action
Neuron
Synapse
Neuron vs node
Training the Network -
Learning
Training the Network - Learning
Backpropagation
 Requires training set (input / output pairs)
 Starts with small random weights

 Error is used to adjust weights (supervised learning)

 Gradient descent on error


Example: Voice Recognition
Task: Learn to discriminate between
two different voices saying “Hello”
Data
 Sources
 Steve
 David
 Format
 Frequency distribution (60 bins)
Network architecture
 Feed forward network
 60 input (one for each frequency bin)
 6 hidden
 2 output (0-1 for “Steve”, 1-0 for “David”)
Presenting the data
Steve

David
Presenting the data
Steve

David
Presenting the data (untrained network)
Steve

0.43

0.26

David

0.73

0.55
Calculate error
Steve

0.43 – 0 = 0.43

0.26 –1 = 0.74

David

0.73 – 1 = 0.27

0.55 – 0 = 0.55
Backprop error and adjust weights
Steve

0.43 – 0 = 0.43

0.26 – 1 = 0.74

1.17

David

0.73 – 1 = 0.27

0.55 – 0 = 0.55

0.82
Repeat process (sweep) for all training
pairs
 Present data
 Calculate error
 Backpropagate error
 Adjust weights

Repeat process multiple times


Presenting the data (trained network)
Steve

0.01

0.99

David

0.99

0.01
Results – Voice Recognition

 Performance of trained network

 Discrimination accuracy between known “Hello”s


 100%

 Discrimination accuracy between new “Hello”’s


 100%
Problems with Expert Systems
Limited domain
Systems are not always
up to date, and don’t
learn
No “common sense”
Experts needed to setup
and maintain system
Legal and Ethical Issues
Who is responsible if the advice is wrong?
 The user?
 The domain expert?
 The knowledge engineer?
 The programmer of the expert system shell?
 The company selling the software?

You might also like