Department of Computer & Information Sciences
Pakistan Institute of Engineering and Applied Sciences
Introduction
CIS-524 Fuzzy Intelligence
Lecture 01
Umar Faiz
http://www.pieas.edu.pk/umarfaiz/cis524
CIS-524 Fuzzy Intelligence
docsity.com
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Introduction-Fuzzy Intelligence-Lecture Slides, Slides of Artificial Intelligence
This lecture was delivered by Dr. Asif Ullah at Pakistan Institute of Engineering and Applied Sciences, Islamabad (PIEAS) for Fuzzy Intelligence course. It includes: Introduction, Complexity, Credibility, Uncertainty, Traditional, Modern, Logic, Probability
Typology: Slides
2011/2012
1 / 34
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download Introduction-Fuzzy Intelligence-Lecture Slides and more Slides Artificial Intelligence in PDF only on Docsity!
Department of Computer & Information SciencesPakistan Institute of Engineering and Applied Sciences
CIS-524 Fuzzy Intelligence
Lecture 01
Umar Faiz
http://www.pieas.edu.pk/umarfaiz/cis
CIS-524 Fuzzy Intelligence
docsity.com
Introduction •
In general practice, we deal with problems in terms ofsystems that are constructed as modelssystems that are constructed as models.
The purpose of constructing models is to understand
h
f
lit
b
it
t
l
some phenomenon of reality, be it natural or man-made.
2
docsity.com
Introduction •
Uncertainty has a pivotal role in any effort tomaximize the usefulness of systems modelsmaximize the usefulness of systems models.
Uncertainty becomes very valuable when considered in
ti
t
th
th
h
t
i ti
f th
t
connection to the other characteristics of the systemsmodels.
The concept of uncertainty in science and mathematicsis manifested by two views
Traditional View
Modern View
4
docsity.com
Introduction •
Traditional View:
According to traditional view science should strive for
According to traditional view, science should strive forcertainty in all its manifestations (precision, specificity,sharpness, consistency, etc) and uncertainty (imprecision, non-
p
,
y,
)
y (
p
,
specificity, vagueness, inconsistency, etc) is considered asunscientific.
Modern View:
According to modern view, uncertainty is considered essential
g
,
y
to science.
5
docsity.com
Fuzzy Logic •
Fuzzy Logic
Fuzzy or multi valued logic was introduced in the 1930s by
Fuzzy, or multi-valued logic, was introduced in the 1930s byJan Lukasiewicz, a Polish philosopher.
While classical logic operates with only two values 1 (true)While classical logic operates with only two values 1 (true) and 0 (false), Lukasiewicz introduced logic that extended therange of truth values to all real numbers in the intervalbetween 0 and 1.
docsity.com
Fuzzy Logic •
In 1965 Lotfi Zadeh, published his famous paper“Fuzzy sets” Zadeh extended the work on possibility
Fuzzy sets. Zadeh extended the work on possibility theory into a formal system of mathematical logic, andintroduced a new concept for applying naturalintroduced a new concept for applying naturallanguage terms.
Thi
l
i
f
ti
d
i
l ti
f
This new logic for representing and manipulating fuzzyterms was called fuzzy logic.
docsity.com
Fuzzy Logic •
The word fuzzy means blurred, fluffy, frayed orindistinctindistinct.
Fuzzy logic is not logic that is fuzzy, but logic that is
d t
d
ib
f
i
used to describe fuzziness.
Fuzziness is deterministic uncertainty.
Fuzziness is connected with the degree to which eventsoccur rather than the likelihood of their occurrence(probability).
For example, the degree to which a person is young is afuzzy event rather than a random event.
docsity.com
Fuzzy Logic •
Fuzzy sets are functions that map each member in a setto a real number in [0 1] to indicate the degree ofto a real number in [0, 1] to indicate the degree ofmembership of that member.
Th
bi
it
f
l
ld d fi iti
The ambiguity of real world definitions
Javaid is OLD
H
"OLD" i
ld? 40
50
60?
H
ow "OLD" is old? 40 years, 50, or 60?
Daud is TALL
How "TALL" is tall? 5 feet, 6 feet, or 7 feet?
Every thing is a matter of degree
The "degrees" of being old or tall can be quantitativelyillustrated using quantified meaning.
docsity.com
Fuzzy Logic Applications •
Replacement of a skilled human operator by a fuzzyrule based systemrule based system
Sendal subway (Hitachi)
Cement kiln (F L Smith)
Cement kiln (F.L. Smith)
Elevator Control (Fujitec, Hitachi, Toshiba)
S
'^
d l
d
d l h li
t
Sugeno s model car and model helicopter
Hirota's robotN
l^
R
C
l (Hi
hi B
d)
N
uclear Reactor Control (Hitachi, Bernard)
Automobile automatic transmission (Nissan, Subaru)B lld
C
l (T
)
B
ulldozer Control (Terano)
Ethanol Production (Filev)
Appliance control – Washing machine, microwave ovens, ricecookers, vacuum cleaners, camcorders, TVs, thermal rugs,heaters
13
docsity.com
Fuzzy Logic Applications •
Replacement of a human expert by a fuzzy logicbased decision making systembased decision making system
Medical – CADIAG
Securities (Yamaichi Hitachi)
Securities (Yamaichi, Hitachi)
Credit Worthiness (Zimmermann)
D
t (Y
H di
i
)
Damage assessment (Yao, Hadipriono)
Fault Diagnosis (Guangzhou)P
d
i
l
i^
(T
k
)
P
roduction planning (Turksen)
14
docsity.com
Neural Networks – Basic Concepts •
A Neural Network generally maps a set of inputs to aset of outputsset of outputs
Number of inputs/outputs is variable
The network itself is composed of an arbitrary numberof nodes with an arbitrary topology
Input 0
Input 1
Input n
Output 0
Output 1
Output m docsity.com
Neural Networks – Basic Concepts •
Neuron vs. Node
Illustration of a biological and artificial neuron (perception)
docsity.com
What can a Neural Net do? •
Has a potential to solve difficult problems currentmethods can not solve well (realistic reasons):methods can not solve well (realistic reasons):
Pattern classification: hand-written characters, facialexpression engine diagnosis etcexpression, engine diagnosis, etc.
Non-linear time series modeling, forecasting: Stock price,utility forecasting, ecg/eeg/emg, speech, etc.
y
g,
g/
g/
g,
p
,
Adaptive control, machine learning: robot arm, autonomousvehicle
Requires massive parallel implementation with opticaldevices, analog ICs.devices, analog ICs.
Performance degrades gracefully when portions of thenetwork are faultynetwork are faulty.
docsity.com
Neural Netwoks •
The study of neural networks started by the publicationof Mc Culloch and Pitts [1943]of Mc Culloch and Pitts [1943].
The single-layer networks, with threshold activation functions,were introduced by Rosenblatt [1959] These types ofwere introduced by Rosenblatt [1959]. These types ofnetworks were called perceptrons.
In the 1960s it was experimentally shown that perceptrons
9
p
y
p
p
could solve many problems, but many problems could not besolved.
These limitations of one-layer perceptron weremathematically shown by Minsky and Papert in their
y
y
y
p
book Perceptron [1969].
20
docsity.com