artificial intelligence - a brief introduction and application examples

25
Ling & Lam 2009 Artificial Intelligence: A Brief Introduction and application examples

Upload: taylor-ling

Post on 18-Nov-2014

4.324 views

Category:

Education


1 download

DESCRIPTION

 

TRANSCRIPT

Page 1: Artificial Intelligence - A Brief Introduction And Application Examples

Lin

g &

Lam

20

09

Art

ific

ial

Inte

llig

en

ce: A

Bri

ef

Intr

od

uct

ion

an

d a

pp

lica

tio

n

ex

am

ple

s

Page 2: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

1

Table of Contents

Lists of Figures and Tables........................................................................................................ 2

Figures ................................................................................................................................... 2

Tables .................................................................................................................................... 2

Chapter 1 Introduction ............................................................................................................... 3

Chapter 2 Applications of AI .................................................................................................... 4

2.1 Games .............................................................................................................................. 4

2.2 Expert System .................................................................................................................. 5

2.3 Intelligent Agent .............................................................................................................. 6

2.4 Simulations ...................................................................................................................... 7

2.5 Computer Vision ............................................................................................................. 8

2.6 Natural Language Processing ........................................................................................ 10

2.7 Machine Learning .......................................................................................................... 12

2.8 Interfaces ....................................................................................................................... 12

2.9 Robotics ......................................................................................................................... 13

2.10 Theorem Proving ......................................................................................................... 14

Chapter 3 Expert System ......................................................................................................... 16

3.1 Introduction ................................................................................................................... 16

3.2 Basic Concepts of Expert Systems ................................................................................ 17

3.4 How Expert Systems Work ........................................................................................... 18

3.5 Expert Systems in Medical Field ................................................................................... 21

3.6 Pros and Cons of Expert Systems .................................................................................. 22

3.7 Conclusion ..................................................................................................................... 22

Chapter 4 Conclusion .............................................................................................................. 24

Reference ................................................................................................................................. 24

Page 3: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

2

Lists of Figures and Tables

Figures

Figure 1 Deep Blue…………………………………………………….………………………5

Figure 2 & Figure 3 Flight Simulation and Auto Racing Simulation………………………….8

Figure 4 Video surveillance system software………………………………………………...10

Figure 5 A map of ALICE's "brain" plots all the words she knows….………………………11

Figure 6 & Figure 7 Android Aiko (Left) and Lisa (Right)…………………………………..14

Figure 8 Process of transferring expertise……………………………………………………18

Figure 9 Main components of expert systems and their interrelationship……………………20

Tables

Table 1 Types of expert systems……………………………………………………………...16

Table 2 Lists of well-known expert systems in medical field………………………………..21

Table 3 Application example in different categories of medical expert systems…………….21

Page 4: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

3

Chapter 1 Introduction

Artificial Intelligence (AI), a term that coined by John McCarthy in his 1955

proposal for the 1956 Dartmouth Conference, is the branch of computer science

concerned with making computers behave like humans, or defined by John McCarthy

himself as “the science and engineering of making intelligent machines, especially

intelligent computer programs.”

After Second World War, a number of people started to work on intelligent

machines independently, where it is believed that that was the time AI research

started. The English mathematician Alan Turing may have been the first where he

gave a lecture on it in 1947. He also may have been the first to decide that AI was best

researched by programming computers rather than by building machines. By the late

1950s, there were many researchers on AI, and most of them were basing their work

on programming computers. Although it has been around 60 years in AI research

where most of the researcher aimed to simulate full human behavior in various

applications, however, up until today, there is still no computers exhibit full artificial

intelligence. Nevertheless, research in AI never stopped where there are still a lot of

different research projects focusing on AI improvement. One of the noticeable

projects is the Artificial Intelligence System (AIS), which is a large scale distributed

computing project by Intelligence Realm, Inc. that involving over 10,000 computers

with the initial goal to recreate the largest brain simulation to date. The AIS project

has successfully simulated over 700 billion neurons as of April 28, 2009.

Although the failure for appearance of AI that fully mimics human behavior,

there are numerous of fields, such as natural resources management, medicine,

military, petroleum industry, just to mention a few, has been make full use of AI to

achieve different goals in a smarter way nowadays. Some of the AI applications have

been proven to give great practical benefits, and despite of the widely use of AI

techniques in software and hardware, the existence of AI in all those products mostly

go unnoticed by many people, which also known as the “AI Effect”. Thus, it is not

hard to observe that AI is so important that it has been a part of life in every

industrialized nation. In the next section, there will be some examples on the

application of AI in different fields or industries.

Page 5: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

4

Chapter 2 Applications of AI

2.1 Games One of the most studied and most interesting areas of AI is the application of

AI in games and simulations. AI has been extensively used many type of games,

including board games like chess, checker; computer games like strategy games

and massively multiplayer online role-playing game (MMORPG). With the

integration of AI, a game will become livelier and more fun to the user as the AI

used is simulating the real environment or human behavior in the game.

AI applied in games somewhat complex if compared to those AI applied in

problem solving system which can perform precise and accurate decision, such as

expert system. AI for game is about the imitation of human behaviour where it has

to be

Creative and Smart (to a certain extent),

Non-repeating behaviour,

Unpredictable but rational decisions,

Emotional influences or personality,

Body language to communicate emotions,

Being integrated in the environment,

so that the user are able to experience the somewhat „human-like‟ response on any

action being taken in the game.

Early research of AI application in game playing (state space search) was done

using common board games such as checkers, chess, and the 15-puzzle. These

games are chosen to be the subject not only because they have limited and well

defined rules, but they also have huge state spaces due to the complexity of the

game which means that a perfect move is almost impossible in limited time. One

of the most successful examples of the game playing AI in the history is the Deep

Blue, a chess-playing computer that derived its playing strength mainly out of

brute force computing power, which is developed by IBM. On May 11, 1997, the

machine won a six-game match by two wins to one with three draws against world

champion Garry Kasparov.

Page 6: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

5

Figure 1 Deep Blue

2.2 Expert System AI programs that achieve expert-level competence in solving problems in task

areas by bringing to bear a body of knowledge about specific tasks are called

expert systems. Often, the term expert systems is reserved for programs whose

knowledge base contains the knowledge used by human experts, in contrast to

knowledge gathered from textbooks or non-experts.

The primary goal of expert systems research is to make expertise available to

decision makers and technicians who need answers quickly. There is always not

enough expertise to go around solving problem at the right place and the right

time. Expert systems can assist supervisors and managers with situation

assessment and long-range planning. Many small systems now exist that bring a

narrow slice of in-depth knowledge to a specific problem, and these provide

evidence that the broader goal is achievable.

Expert systems have been widely used in a number of industries; one of the

examples that utilize it extensively is the aviation industry. The scheduling flights

based on economics, environmental, regulatory requirements and airway traffic

parameters are extremely complex. Any mistakes in any one of the steps involved

in a flight is extremely costly. Typically, expert systems that can be found in

aviation industry include:

Page 7: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

6

The Aviation Expert System

This is an expert system that used to clarify psychological

assessment issues in the field of aviation.

General Aviation Pilot Advisor and Training System (GAPATS)

GAPATS is a computerized airborne expert system that uses fuzzy

logic to infer the flight mode of an aircraft from:

o sensed flight parameters

o an embedded knowledge base, and

o pilot inputs

which the data will then used to assess the pilot‟s flying performance

and issue recommendations for pilot actions.

Aircraft Maintenance Expert Systems (AMES)

AMES has been used since the early 1990's. Manual procedures

around aircraft maintenance are very strenuous and time consuming.

Diagnosis of aircraft malfunctions is an ideal candidate for an expert

system to assist in the diagnosis of aircraft problems.

Anti-G Fighter Pilot System

The high maneuverability of modern jet fighters often subjects the

pilots to high Gz acceleration. One of the adverse effects of Gz

acceleration is loss of consciousness. The Anti-G Fighter Pilot System

presents an alternative to the current protection pressure mask and

pressurized G-suit. The system used expert knowledge and pilots'

anthropometric and physiologic data to generate control schedules of

the G-suit and mask pressures of jet fighter pilots.

2.3 Intelligent Agent In AI, an intelligent agent (IA) is an autonomous entity which observes

and acts upon an environment and directs its activity towards achieving goals.

Intelligent agents may also learn or use knowledge to achieve their goals.

Page 8: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

7

They may be very simple or very complex; for example, a reflex machine such

as a thermostat is an intelligent agent.

Based on the intelligent agent‟s degree of perceived intelligence and

capability, Russell & Norvig categorize agents into five classes:

i. simple reflex agents

ii. model-based reflex agents

iii. goal-based agents

iv. utility-based agents

v. learning agents.

However, intelligent agents nowadays are normally gathered in a

hierarchical structure containing many “sub-agents” in order to perform more

advance tasks. Intelligent sub-agents process and perform lower level

functions. Taken together, the intelligent agent and sub-agents create a

complete system that able to accomplish difficult tasks or goals with behaviors

and responses that display a form of intelligence.

In the world of e-commerce, intelligent agents known as shopping bots are

used by consumers to search for product and pricing information on the Web.

Each shopping bot operates differently, depending on the business model used

by its operator. Famous internet search engine like Google, Yahoo!, and

Ask.com are also utilizing this technology to perform information searching

for the internet user.

2.4 Simulations Besides the AI used in game playing with the purpose of having fun,

simulations nowadays has heavily written code for AI to simulate the most

realistic feedback and response for any action by the user; to let the user feels

that he is placed in that particular environment or situation. This is very

important for the industries that utilizing simulations, such as the aviation

industry and auto racing because any slight mistake in a flight or in a race is

very costly, even death. Thus, simulation is used to train the user familiar with

some events that rarely happen but able to respond swiftly if that particular

Page 9: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

8

event really happens. In a simulator, every input of the user will get responded

by the most accurate feedback to ensure that the user is able to experience it

before he/she go for the real thing, especially in auto racing industry.

The application of artificial intelligence in simulators is proving to be very

useful for the aviation industry. Airplane simulators are using artificial

intelligence in order to process the data taken from simulated flights. Other

than simulated flying, there is also simulated aircraft warfare. The computers

are able to come up with the best success scenarios in these situations. The

computers can also create strategies based on the placement, size, speed, and

strength of the forces and counter forces. Pilots may be given assistance in the

air during combat by computers. The artificial intelligent programs can sort

the information and provide the pilot with the best possible maneuvers, not to

mention getting rid of certain maneuvers that would be impossible for a

sentient being to perform. Multiple aircraft are needed to get good

approximations for some calculations so computer simulated pilots are used to

gather data. These computer simulated pilots are also used to train future air

traffic controllers.

Figure 2 & Figure 3 Flight Simulation and Auto Racing Simulation

2.5 Computer Vision

Computer vision is the branch of artificial intelligence that focuses on

providing computers with the functions typical of human vision. Computer

Page 10: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

9

vision has been applied in several fields such as industrial automation,

robotics, biomedicine, and satellite observation of Earth. In the field of

industrial automation alone, its applications include guidance for robots to

correctly pick up and place manufactured parts, nondestructive quality and

integrity inspection, and on-line measurements.

When the famous dedicated computer-vision system, the Massively

Parallel Processor (MPP) designed at the Goddard Space Flight Center in 1983,

it does not received good response due to the complexity and very high price.

However, with the advancement in manufacturing technology, the availability

of affordable hardware and software has opened the way for new, pervasive

applications of computer vision. Computer-vision systems have one factor in

common. They tend to be human-centered; that is, either humans are the

targets of the vision system or they wander about wearing small cameras, or

sometimes both. Nowadays, computer-vision systems have been used in quite

a number of applications, including:

o human-computer interfaces (HCIs), the links between computers and

their users

o augmented perception, tools that increase normal perception

capabilities of humans

o automatic media interpretation, which provides an understanding of the

content of modern digital media, such as videos and movies, without

the need for human intervention or annotation

o video surveillance and biometrics.

Although the success of the usage of computer vision in our daily life, to

date, computer vision systems still unable to emulate the full capabilities of

the human visual system. The human eye-brain combination still proved to be

the best which is able to categorize previously unseen objects with ease, using

background knowledge and context. Nevertheless, computer vision has helped

to solve a lot of problems and difficulties in human‟s life; hence it is

Page 11: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

10

foreseeable that the usage of computer vision will continue to grow to improve

humans‟ life.

Figure 4 Video surveillance system software

2.6 Natural Language Processing Natural Language Processing (NLP) is both a modern computational

technology and a method of investigating and evaluating claims about human

(natural) language itself. NLP uses computers to process written and spoken

language for some practical, useful, purpose: to translate languages, to get

information from the database to answer an enquiry, to carry on conversations

with machines, so as to get advice about, say, pensions and so on. And these are

only examples of major types of NLP. There is also a huge range of lesser but

interesting applications, e.g. getting a computer to decide if one newspaper story

has been rewritten from another or not. Hence, NLP is not simply applications but

the core technical methods and theories that the major tasks above divide up into,

such as Machine Learning techniques, which is automating the construction and

adaptation of machine dictionaries, modeling human agents' beliefs and desires

etc. Artificial Intelligence is an essential component of NLP if the computers have

to engage in realistic conversations with human.

Page 12: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

11

One of the modern AI research in NLP is the research in chatterbot. A

chatterbot (or chatbot) is a type of conversational agent, a computer program

designed to simulate an intelligent conversation with one or more human users via

auditory or textual methods. Also known as Artificial Conversational Entity

(ACE) and, though many appear to be intelligently interpreting the human input

prior to providing a response, most of the chatterbots simply look for keywords

within the input and reply with the most matching keywords or the most similar

wording pattern from a local database. The classic and early famous chatterbots

are ELIZA (1966) and PARRY (1972). More recent programs are Racter, Verbots,

A.L.I.C.E., SmarterChild, and ELLA. With the growing number of research in

chatterbots, the initial purpose of creating chatterbots has been expanded to many

other usages, for example story „writing‟ by Racter.

Some of the organizations are already beginning to implement a so-called

Automated Conversational Systems which is a system that evolved from the

original designs of the first widely used chatbots. In the UK, large commercial

entities such as Lloyds TSB, Royal Bank of Scotland, Renault, Citroën and One

Railway are already utilizing Virtual Assistants to reduce expenditures on Call

Centres and provide a first point of contact that can inform the user exactly of

points of interest, provide support, capture data from the user and promote

products for sale.

Figure 5 A map of ALICE's "brain" plots all the words she knows.

Page 13: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

12

2.7 Machine Learning Machine learning is a scientific discipline that is concerned with the design

and development of algorithms that allow computers to learn based on data, such

as from sensor data or databases. A major focus of machine learning research is to

automatically learn from past examples and to detect hard-to-discern patterns from

large, noisy or complex data sets.

Expert systems and data mining programs are the most common applications

for improving algorithms through the use of machine learning. Among the most

common approaches are the use of artificial neural networks (weighted decision

paths) and genetic algorithms (algorithms “bred” and culled to produce

successively fitter programs).

Due to the nature of machine learning, its‟ capability is particularly well-suited

to medical applications, especially those that depend on complex proteomic and

genomic measurements. As a result, machine learning is frequently used in cancer

diagnosis and detection. More recently machine learning has been applied to

cancer prognosis and prediction. This latter approach is particularly interesting as

it is part of a growing trend towards personalized, predictive medicine. Among the

better designed and validated studies it is clear that machine learning methods can

be used to substantially (15–25%) improve the accuracy of predicting cancer

susceptibility, recurrence and mortality. At a more fundamental level, it is also

evident that machine learning is also helping to improve basic understanding of

cancer development and progression.

A trend can be seen that machine learning has been used quite often in a lot of

different systems, be it a diagnostic system or education system, due to the long

term benefit. The success of machine learning programs suggests the existence of

a set of general learning principles that will allow the construction of programs

with the ability to learn in realistic domains.

2.8 Interfaces Even the most sophisticated and powerful system will be next to useless

without an effective user interface. A good and user-friendly interfaces that

controls a complex machine is must-equipped software to allow the interaction

between the user and the machine. However, with the rapid advancement in

Page 14: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

13

technology, the demand in interfaces become higher and higher. Most of the users

already started to ask for a smarter interface to learn and remember their

preferences, instead of having the user input the same preference everytime and

thus, interface agents are started to get developed due to the demand.

Interface agents are computer programs that employ Artificial Intelligence

methods to provide active assistance to a user of a particular computer application.

The metaphor used is that of a personal assistant who is collaborating with the

user in the same work environment. The assistant becomes gradually more

effective as it learns the user's interests, habits and preferences.

One of the examples of commercialized interface agent has been used in

Microsoft Windows Vista called SuperFetch, a feature that predicts which

applications are used when, then pre-loads them so that they're instantly available.

„Microsoft Research contributed to the SuperFetch effort, a feature within Vista

that predicts which applications are used when, then pre-loads them so that they're

instantly available.‟ „As part of a long term set of projects, we want to teach the

computer to learn from users to make the machine more proactive,‟ says Eric

Horvitz, a principal researcher with Microsoft's R&D as well as the president-

elect of the American Association for Artificial Intelligence. „We want to use the

system's idle time to make things punchier.‟

Therefore, it is not surprising that the interface agent become part of the

important components, no matter in the hardware or in the software, due to the

high demand of human.

2.9 Robotics

Robots have become common in many industries. They are often given jobs

that are considered dangerous to humans. Robots have proven effective in jobs

that are very repetitive which may lead to mistakes or accidents due to a lapse in

concentration and other jobs which humans may find degrading. General Motors

uses around 16,000 robots for tasks such as painting, welding, and assembly.

Japan is the leader in using and producing robots in the world. In 1995, 700,000

robots were in use worldwide; over 500,000 of which were from Japan. Thus,

Page 15: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

14

from the number itself, it can be seen that it is very important to have AI in the

robots to keep the productivity at a constant level.

Besides the application of AI in heavy industries robots, AI also part of the

crucial components in human-like robots such as famous Android Aiko and

ASIMO. Both of the robots have the ability to recognize speech, voice, faces,

motion, objects and also learn simple thing with the AI implemented in the

software that operating them. Without AI, a robot will be doing a same thing

everytime as programmed.

The latest breakthrough in robotics AI happens when a relatively new

company called AI Robotics based in Kobe, Japan has developed a female robot

called „Lisa – The Perfect Woman‟ which equipped with RKS, “Recognition Krax

System”, which allows for vocal, tactile and visual recognition. Lisa is able to

recognize objects and persons and she can even differentiate between roses and

tulips for example. Claimed by the creators, Lisa can cook a meal based on what

is in the fridge using visual recognition. She also can go shopping, do household

work or give a hydraulic massage, and she can also play chess and video games

and even learn to do certain sports.

Figure 6 & Figure 7 Android Aiko (Left) and Lisa (Right)

2.10 Theorem Proving

Automated theorem proving is appealing due to the rigor and generality of

logic. Because it is a formal system, logic lends itself to automation. A wide

variety of problems can be attacked by representing the problem description and

relevant background information as logical axioms and treating problem instances

as theorems to be proved. This insight is the basis of work in automatic theorem

proving and mathematical reasoning systems.

Page 16: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

15

Unfortunately, early efforts at writing theorem provers failed to develop a

system that could consistently solve complicated problems. This was due to the

ability of any reasonably complex logical system to generate an infinite number of

provable theorems: without powerful techniques (heuristics) to guide their search,

automated theorem provers proved large numbers of irrelevant theorems before

stumbling onto the correct one. In response to this inefficiency, many argue that

purely formal, syntactic methods of guiding search are inherently incapable of

handling such a huge space and that the only alternative is to rely on the informal,

ad hoc strategies that humans seem to use in solving problems. This is the

approach underlying the development of expert systems (Chapter 8), and it has

proved to be a fruitful one.

Still, the appeal of reasoning based in formal mathematical logic is too strong

to ignore. Many important problems such as the design and verification of logic

circuits, verification of the correctness of computer programs, and control of

complex systems seem to respond to such an approach. In addition, the theorem-

proving community has enjoyed success in devising powerful solution heuristics

that rely solely on an evaluation of the syntactic form of a logical expression, and

as a result, reducing the complexity of the search space without resorting to the ad

hoc techniques used by most human problem solvers.

Another reason for the continued interest in automatic theorem provers is the

realization that such a system does not have to be capable of independently

solving extremely complex problems without human assistance. Many modern

theorem provers function as intelligent assistants, letting humans perform the

more demanding tasks of decomposing a large problem into subproblems and

devising heuristics for searching the space of possible proofs. The theorem prover

then performs the simpler but still demanding task of proving lemmas, verifying

smaller conjectures, and completing the formal aspects of a proof outlined by its

human associate.

Page 17: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

16

Chapter 3 Expert System

3.1 Introduction

As mentioned previously, an expert system is actually an advisory system that

embedded with human expertise which mainly used in solving particular types of

problems. Expert systems were developed in the mid-1960s and have begun to

emerge in the following decade. By 1980s, it has also started to be implemented in

commercial field other than academic applications. Expert systems have being the

most common applied AI technology up to nowadays.

Expert systems can be used in various problem areas, such as interpretation

systems, prediction systems, diagnostic systems, design systems, developing plans to

achieve goal, comparing observations, debugging, repairing some diagnosed

problems, instructing or correcting student performance and controlling the system

behavior.

Below are several types of expert systems:

Categories Description

Rule-based expert systems It is normally implemented in commercial field, its

knowledge is presented as production rules (i.e. IF

some conditions, THEN action).

Frame-based expert systems The knowledge is presented as frames, which the

knowledge of a particular object is organized in a

special hierarchical structure.

Hybrid expert systems There includes a combination of several knowledge

representation approaches in this systems.

Model-based expert systems Model being used in this system as reference in

comparing with experimental subject.

Systems classified by their

nature

This system will classify the nature of the problem

and solve it by retrieving data according to the

class of the problem.

Ready-made expert systems Computer systems that are designed based on

general use instead of the needs of particular users.

Real-time expert systems It has to response as fast as possible by the time it

is needed.

Table 1 Types of expert systems

Page 18: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

17

The remainder of this assignment will be introducing the fundamentals of

expert systems, which includes the basic ideas of expert systems and how it works.

We will also briefly introduce the application of expert systems in the medical field.

3.2 Basic Concepts of Expert Systems

In consequence of the fast growing technologies nowadays, experts are highly

on demand in this competitive world. Limited experts in various fields such as

engineering, has become a problem for them to stay in pace with the development. In

response to this problem, knowledge lies within experts is needed to be captured.

Expert systems in this case act as an advisory system or even can be used by

experts as knowledgeable assistants. In building such expert systems, the knowledge

of experts will be transferred into a computer system. It can then provide some

advices even explanations to non-experts or novices. The basic concepts of expert

systems include: expertise, experts, transferring expertise and explanation capability.

Expertise

Expertise, normally defined as a high level of knowledge or skill, but in expert

systems we define it as a specific knowledge or skill acquired from training, reading

and experience. It is the main element that is to be implemented in the expert systems,

which enables non-experts or novices to solve particular type of problems.

Experts

The so-called expert is the person who possessed of expertise in their field and

playing the most important role in problem solving and decision making. They are the

main knowledge resources in building expert systems. To mimic the human experts,

expert systems need to have the ability in solving problems, learn from experience,

restructure knowledge and rebuild the rules where necessary.

Transferring expertise

Transferring expertise is the main objective of the expert systems, which

transfer from expert to computer system and then to non-experts or novices. The

transferring process basically involves knowledge acquisition, representation,

Page 19: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

18

inferencing and transfer to the user. Knowledge acquisition is the process that

transfers the knowledge of experts to the computer system, which then stored in the

system and represented in the computer. And knowledge inferencing is a unique

feature of the expert systems which enables the systems to reason. After inferencing

knowledge, finally it will transfer the solution to the user based on the rules and facts

regarding to the problems.

Explanation capability

Expert systems are distinct from the conventional computer systems, which it

is capable in explaining its advices or operations. This feature allows users to

understand more on the advices provided and also enables the system to justify its

own reasoning.

3.4 How Expert Systems Work

Common components in Expert Systems

There are few common components that might exist in the expert systems:

Knowledge acquisition subsystem

It is responsible in constructing and expanding the knowledge base by

transferring the knowledge from all possible sources such as human experts,

Figure 8 Process of transferring expertise

Human

Experts

Knowledge

Engineer User

Kn

ow

led

ge

Base

Inference

Engine

Data, problems, questions

Knowledge, concepts, solutions

Page 20: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

19

books, databases and graphical resources to the computer system. A

knowledge engineer might be playing an important role in this process.

Knowledge base

Typically, the knowledge base is separated into facts and rules. Facts are

something like problem situations and theories underneath, whereby rules are

used to manage suitable knowledge in solving particular problems.

Inference engine

This component is the reasoning tool of the expert systems. It interprets on the

information from knowledge base and blackboard, and provides a direction to

the appropriate system‟s knowledge.

Blackboard

It is also known as the workplace of the expert systems, which mainly used in

recording the specified problem (input) and also the intermediate hypotheses

and decisions.

User interface

Expert systems are user-friendly, which its user interface provides a platform

for user to communicate with the system in natural language.

Explanation subsystem

This subsystem provides the explanation capability to the expert systems.

Knowledge refining system

Refinement is essential to build a good expert system. With refinement, expert

systems able to check on its own performance, make improvements and learn

from experience.

Human elements involved

The main human elements that involved in expert systems are as follows:

Experts

Knowledge engineer

Basically, a knowledge engineer is responsible in interacting with human

experts to build the knowledge base. In building the expert systems, they may

cooperate with other computerized systems to integrate it.

User

Page 21: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

20

How it works?

Basically, the operation of the expert systems can be categorized into 2 parts:

development and consultation.

In the first stage, development, activities that involved includes constructing

the knowledge base, inference engine, blackboard, explanation facility and any other

necessary software. These activities may be complex; therefore a tool called ES shell

has been used to speed up the process. The ES shell includes all basic components of

expert systems except the knowledge. Improvement can be done to the systems by

rapid prototyping during their development. Rapid prototyping will represent the

knowledge acquired in a better manner that allows quick inference process.

After completed the construction stage, the system will be tested and

validated, and then comes to the consultation part. Expert systems are well designed

to be user-friendly with intuitive features. Hence, the user can easily input their

problems and get advices and explanation in the consultation environment. During the

consultation process, more questions may be asked and answered to reach a

conclusion. The inference engine involve in the reasoning process whereby

explanation facility provides explanations to the user.

Figure 9 Main components of expert systems and their interrelationship

Use

r In

terf

ace

Bla

ckb

oard

Inference

Engine K

now

led

ge

Acq

uir

ing S

ub

syst

em

Knowledge

Base

Explanation

Subsystem

User

Page 22: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

21

3.5 Expert Systems in Medical Field

Concerning medicine, there are a number of expert systems for different usage

have been developed. Some of the well-known examples are as follows:

Expert systems Description

MYCIN For diagnosing bacteria that cause severe infections and

recommending antibiotic dosage according to different

patient‟s body weight

HELP A complete knowledge based hospital information

system

deDombal`s Leeds

Abdominal Pain System

For acute abdominal pain

Table 2 Lists of well-known expert systems in medical field

Other than the above common medical expert systems, there is more expert

systems have been developed recent years in assisting medical works. The

development includes in:

Acute care systems,

Decision support systems,

Education systems,

Quality assurance and administration,

Medical imaging,

Drug administration,

Laboratory systems. (Coiera, 1997)

Categories Example application

Acute care systems Coronary care admission, giving advices on the

management of chest pain patients in the emergency room.

Decision support

systems

Typical example is HELP system that has been mentioned

above.

Education systems Delivering knowledge on how to reduce risk of cancer.

Quality assurance Monitoring patient clinical data for potential adverse drug

effect.

Medical imaging Automatic interpretation of medical imaging data such as

Cardiac SPECT data.

Laboratory systems Haematology analyzer

Table 3 Application example in different categories of medical expert systems

Page 23: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

22

3.6 Pros and Cons of Expert Systems

Advantages of expert systems

Expert systems have brought numerous benefits to the users:

Productivity will increase as expert systems work much faster than human.

And the output will also increase due to the reduction of workers needed

which in turn reduce in costs.

Scarce expertise can be captured.

Minimize the employees training costs.

Expert systems can be made in many copies where eliminate the needs of

human experts to travel around and increase the accessibility of expertise.

Expert systems can integrate knowledge from several experts.

Expert systems can operate in hazardous environments and they will not affect

by temper or tiredness because they do not have feelings.

Limitations of expert systems

Beside benefits stated above, there are also some problems and limitations in dealing

with expert systems:

It is not easy to gather knowledge for the computer systems.

Situation assessment may vary from different experts in different conditions.

Expert systems may make mistakes.

There is no common sense being used in making decision.

The construction of expert systems is complex and expensive.

Expert systems will only work well in a narrow domain.

Expert systems have no flexibility and ability to adapt to changing

environment or gaining new knowledge without reprogramming.

3.7 Conclusion

Expert systems are very beneficial computer systems which able to give

crucial advices or opinions to non-experts by using their acquired knowledge. These

systems are definitely not replacing the current available human experts but are

assisting experts in dealing with their job or decision making process. Although

Page 24: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

23

expert systems are having a very wide range of applications in many fields, there are

still some limitations and problems occur as we discussed in the previous section.

Following the technology and market trend, hard work must be done on expert

systems in solving these limitations for the systems to stay competitive. The future of

expert systems may include the following: increasing system learning capabilities,

using multiple sources of expertise, improving reasoning capabilities and combining

few expert systems working together.

Besides that, expert systems may also cooperate with other technology such as

fuzzy logic, robotics, neural network and so forth. Neural network in this case could

help easing the knowledge acquisition task in the construction of expert systems,

which can cut down the cost and time in employing a knowledge engineer.

Therefore, current expert systems still can be potentially improved. Various

research topics that have great impact in improving expert systems are under

investigation.

Page 25: Artificial Intelligence - A Brief Introduction And Application Examples

Universiti Tunku Abdul Rahman

24

Chapter 4 Conclusion

We have attempted to define artificial intelligence through discussion of its

major areas of research and application. This has shown that AI can be applied in

intelligent problem solving, planning, and communication skills to a wide range of

practical problems. Some of the common features in the application of AI in every

fields including:

1. The use of computers to do reasoning, pattern recognition, learning, or

some other form of inference.

2. The use of large amounts of domain-specific knowledge in solving

problems. This is the basis of expert systems.

3. The purpose of using AI mainly to increase the productivity and decrease

the workload of a human.

As a conclusion, the application of AI has become very wide, and it is

expected to expand even wider to other fields. It is predictable that in the future,

some of the AI application will get combined and builds up an even more

powerful and stronger AI system.

Reference Turban, Efraim. (1992), Expert systems and applied artificial intelligence. Macmillan.

Russell, Stuart J.; Norvig, Peter (2003), Artificial Intelligence: A Modern Approach (2nd ed.),

Upper Saddle River, NJ: Prentice Hall.

Wikipedia – Application of Artificial Intelligence .(2009), Retrieved June 20, 2009, from

Wikipedia website: http://en.wikipedia.org/wiki/Applications_of_artificial_intelligence

Machine learning. (2009). In Encyclopædia Britannica. Retrieved June 24, 2009, from

Encyclopædia Britannica Online:

http://www.britannica.com/EBchecked/topic/1116194/machine-learning

AI Overview. (2008). Retrieved June 19, 2009 from AAAI.org website:

http://www.aaai.org/AITopics/pmwiki/pmwiki.php/AITopics/AIOverview

Microsoft Predicts The Future With Vista's SuperFetch. (2007). Retrieved June23 2009, from

Information Week website:

http://www.informationweek.com/news/windows/showArticle.jhtml?articleID=196902178