Está en la página 1de 36

Intelligent System

Haruki Ueno

Professor
Human-Machine Symbiosis Research
Intelligent Systems Research Division
National Institute of Informatics (NII)

Professor
Department of Informatics
School of Mathematical and Physical Science
The Graduate University of Advanced Studies (Sokendai)
(総合研究大学院大学)
Contents
AI: Concepts, something on history, examples and issues

Knowledge representation models

Expert System

Object model 1 and Software Engineering

Object model 2 and Program Understanding

Program Understanding for Learning Programming: ALPUS & Intellitutor

Symbiotic Information System and Symbiotic robots: HARIS & ROBOTA

Internet-based Distance Learning for Higher Education and WebLS

Engineering Ethics: Concepts and issues


Historical overview of AI

1956 Idea of AI was introduced

l Inference-oriented approaches were tried


in game domain: Checker, Chess
1960’s

1965 DENDRAL project began at Stanford


The first Knowledge-based approach was introduced.
“In the knowledge lies the power” Ed. Feigenbaum

1970’s Knowledge-based approaches were tried in a various


domains.

Born of Knowledge-Based Expert Systems


I.e., KB + IE
Rules, Frames, Semantic Networks

Trials were done in laboratories of US universities.

1980-82 Knowledge representation languages appeared


EMYCIN, EXPERT, OPS5, AGE, COMEX, etc
Historical overview of AI (cont.)

1980’s Industrialization of AI AI-REUM was developed.


LISP machines provided by US ventures.
COMEX was available through
Common LISP proposed for SUN workstations. Fujitu group.
1982 Fifth Generation Project started in Japan.
FMS(ZERO) developed
PROLOG was chosen for a common language.
Concept of Object model introduced
1983-88 AI fever was occurred in Japan.
Intellitutor project started.

1990’s Data-based approaches were tried.


Data mining, sensor-based robotics.
HARIS project started.
Universal knowledge modeling: Ontology
Softcomputing: Fuzzy, Neuro, GA, etc. PREPARE project started (IPA)
1994 Internet was introduced world wide
Web-mining NII established in 2000
Symbiotic Information System
Internet-based Distance Learning systems
Project started
2000’s Semantic Web Department of Informatics started
2004 Semantic computing initiative in Japan
What is Artificial Intelligence?
http://aisig.uum.edu.my/azizi/TN6023/chapter1_1.cfm

AI is a branch of computer science which concern the study and creation of computer
system that exhibit some form of intelligence;
AI is a system that:
can learn new concepts and tasks
can reason and draw useful conclusions about the world around us
can understand a natural language
can perceive and comprehend a visual scene
perform other types of task that require human types of intelligence

Some definitions of AI from different researchers:


AI is the study of ways to make computers be intelligent
It is an attempt to make computer respond like human being
Behavior by a machine that if performed by a human being, would be called intelligent
AI is the study of how to make computer do things at which people are better

Most experts agree that AI is concerned with 2 basic ideas:


1. It involves studying the thought processes of humans
2. It deals with representing those processes via machines (computers, robot etc.)
What is Artificial Intelligence? (cont)
http://aisig.uum.edu.my/azizi/TN6023/chapter1_1.cfm

The objectives of AI are:


1. to make machine smarter
2. to understand what intelligence is
3. to make machine more useful

Meaning of intelligence behavior.


Several abilities are considered signs of intelligence:
learn or understand from experience
make sense out of ambiguous or contradicting messages
respond quickly and successfully to a new situation
use reasons in solving problems and conduct it effectively
apply knowledge to manipulate environment
acquire and apply knowledge
think and reason
recognize the relative importance of different elements in a situation
Natural Intelligence vs Artificial Intelligence
http://aisig.uum.edu.my/azizi/TN6023/chapter1_3.cfm

Some advantages of natural intelligence (NI) over AI:

NI is creative, whereas AI is rather uninspired. In AI, tailored knowledge must be built


into a carefully constructed system.

NI enables people to benefit by the use of sensory experience directly, whereas AI


system must work with symbolic input.

Perhaps the most importantly, human reasoning is able to make use at all times of a
wide context of experience and to deal an individual problem, whereas AI systems
typically gain their power by having a very narrow focus.

The advantages of natural intelligence over AI result in many limitations of Expert


System.
Natural Intelligence vs Artificial Intelligence
http://aisig.uum.edu.my/azizi/TN6023/chapter1_3.cfm

Some advantages of AI over natural intelligence:

AI is permanent whereas natural intelligence is perishable.

AI offers ease of duplication whereas transferring a body of knowledge from one


person to another is a lengthy process.

AI can be less expensive than natural intelligence.

AI, being computer technology, is consistent and thorough whereas natural intelligence
is erratic (not consistent)

AI can de documented whereas natural intelligence is difficult to reproduce because


the person that makes the decision might forget at a later date why he made such
decision.
Basic Q&A by Jhon McCarthy

Q. What is artificial intelligence?

A. It is the science and engineering of making intelligent machines, especially


intelligent computer programs. It is related to the similar task of using computers
to understand human intelligence, but AI does not have to confine itself to
methods that are biologically observable.

Q. What is intelligence?

A. Intelligence is the computational part of the ability to achieve goals in the world.
Varying kinds and degrees of intelligence occur in people, many animals and
some machines.
Basic Q&A by Jhon McCarthy (cont.)

Q. Does AI aim at human-level intelligence?

A. Yes. The ultimate effort is to make computer programs that can solve problems and
achieve goals in the world as well as humans. However, many people involved in
particular research areas are much less ambitious.

Q. How far is AI from reaching human-level intelligence? When will it happen?

A. A few people think that human-level intelligence can be achieved by writing large
numbers of programs of the kind people are now writing and assembling vast
knowledge bases of facts in the languages now used for expressing knowledge.

However, most AI researchers believe that new fundamental ideas are required,
and therefore it cannot be predicted when human level intelligence will be achieved.
What is Artificial Intelligence? By Ueno

Research on Realizing an Intelligent System based on Natural Intelligence.

Understanding of Human Intelligence is the base.

Three elements for realizing an intelligent system:

- Understanding of natural intelligence


- Technology to realizing an intelligent system
- Domain knowledge
What is Intelligence?

1. Ability of Learning

Learning from experience.


- Appending/increasing of knowledge
- Systematizing/structuring of knowledge
- Discovery of ‘rules of the thumb”
- Discovery of knowledge from the data: Data mining

2. Ability of reasoning
reasoning, inference, guessing

Extracting semantics from the data


Judging from the situation
Usage of the knowledge

Simplified image of knowledge-based reasoning:

Reasoning
Data Conclusion

Knowledge
Two aspects of AI research

1. Investigation of the principle of human intelligence

- Cognitive Science, basic research of AI

Intelligence, Reasoning, memory, knowledge, learning, common sense, etc.

2. Make a computer more useful

- Knowledge Engineering, Expert System, application research of AI

Computation Understanding
Database Decision making
Computer control Consultation
Discovery
Backward reasoning

Combined
reasoning

Forward reasoning
Characteristics of Production System
Strong points:
Each rule is easy to define and understanding
Easy in appending and deleting
Simple reasoning control mechanism

Weak points:
Hard to understand the relationships among the Rules
Hard to maintain a large knowledge base
Slow inference speed
Inflexible in knowledge modeling and reasoning control
Different from human’s knowledge structure

Production System is useful for a small knowledge system.

Large knowledge system should be realized by


Frame-based system.
Related fields to AI

Artificial Intelligence (AI)


Research on realizing an intelligent system
A branch of Computer Science

Cognitive Science (CS)


Research on basic mechanism of intelligence
A branch of Psychology

Knowledge Engineering (KE)


Research on knowledge-based systems
Application of AI

Expert System (ES)


Research on knowledge-based problem solving systems

Fuzzy System
Research on Fuzzy logic and its application

Neuro System
Research on Recognition system by means of neuron’s mechanism

Softcomputing
Research on computation-based intelligence, including Fuzzy, Neuro, Genetic algorism
Contributions and Limitations of DENDRAL

También podría gustarte