Materialy/06/Lecture2- ICM Artificial Intelligence
Download
Report
Transcript Materialy/06/Lecture2- ICM Artificial Intelligence
Slovak University of Technology
Faculty of Material Science and Technology in Trnava
Intelligent Control
Methods
Lecture 2: Artificial Intelligence
Concept „Artificial Intelligence“
A competence of machines to think
A science branch
1950:
Turing: Computing Machinery and
Intelligence:
„Can machines think?“
2
Intelligence:
A competence to respond to new situation threw
an activity correction.
A bee?
An
adaptive controller?
Vague answers and definitions.
„Intelligence“ and „thinking“ are philosophical
(not technical) categories.
3
Intelligence (a technical point of view):
A complex of technical and program tools,
techniques and methods for creation of systems,
which are able to master tasks, which need
„natural“ intelligence.
Analogy: Bird´s and man´s flying
Machines do not (and never will) think like a
man (determined with a society and with his
material, physical, intellectual and emotional
needs), but they can simulate the external
manifestations of thinking by their own tools.
4
History of AI:
Classical period (1950 – 1965)
Romantic period (1965 – 1975)
Review of tools and methods
Methods combination
Period of expert systems (1985 – 1995)
Universal problem solvers
Special program languages
Modern period (1975 – 1985)
Declarative knowledge representation
Problem solution methods: theorem proving, graph searching
Expert systems
Period of AI integration into information technologies
(1995 - ...)
5
Problem core of AI:
Information recognition, analysis and interpretation
Written,
spoken, acoustic, graphical, tactile, ...
Knowledge representation
Knowledge
formalization into the form, in which they can
be effective stored and manipulated
Problems solution (inference)
Problems of man-machine communication,
robotics, expert systems, machine learning,
neuronal nets, genetic algorithms... )
6
Technical system is intelligent, if it:
Perceives the environment
Predicts the environment handling
Often without the environment model
Uses obtained information for problem solution
Plans the solution implementation
Communicates with another intelligent systems (human
or technical)
Learns
and understands, what is important
From experiences
From generalization
Adapts its behavior
7
AI perspectives:
Methods, techniques and tools of AI are
integrated into another ICT
(pattern
and speech recognition, ...)
The quality of AI system depends primarily on
used knowledge (and only secondarily on
knowledge manipulation mechanisms)
Direction at specialized knowledge
acquisition, representation and exploitation.
8