Download Artificial Intelligence - Computer Science Department

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Knowledge representation and reasoning wikipedia , lookup

Artificial intelligence in video games wikipedia , lookup

AI winter wikipedia , lookup

Embodied cognitive science wikipedia , lookup

Technological singularity wikipedia , lookup

Philosophy of artificial intelligence wikipedia , lookup

History of artificial intelligence wikipedia , lookup

Ethics of artificial intelligence wikipedia , lookup

Intelligence explosion wikipedia , lookup

Existential risk from artificial general intelligence wikipedia , lookup

Transcript
Introduction to Artificial Intelligence
Introduction
What is Artificial Intelligence?
Notes
It is a young area of science (1956)
Its goals are what we consider Intelligent behaviour
There are many approaches from different points of view
It has received influence from very diverse areas (Philosophy,
Mathematics, Psychology, Biology)
Involves many areas of work with generic goals (learning, perception,
problem solving . . .) and specific goals (chess, diagnosis of diseases,
driving cars, ...)
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
1 / 20
Introduction
A definition of Artificial Intelligence
Notes
Systems that act like humans
Systems that think like humans
“The study of how to make computers
do things at which, at the moment, people are better” (Rich & Knight, 1991)
“The exciting new effort to make computers think ... machines with minds, in the
full and literal sense” (Haugeland, 1985)
“Computational intelligence is the study
of the design of intelligent agents” (Poole et al, 1998)
“The study of mental faculties through
the use of computational models” (Charniak & McDermott, 1985)
Systems that act rationally
Systems that think rationally
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
2 / 20
Introduction
Systems that act like humans
Notes
The model is the human being, the goal is to build a system that can
pass as human.
Turing’s Test: If a system pass the test is intelligent (?)
Abilities: Natural language processing, knowledge representation,
reasoning, learning.
It is not the goal of AI to pass the test
The interaction between programs and people brings the necessity to
make programs that act like humans.
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
3 / 20
Introduction to Artificial Intelligence
Introduction
Turing’s test
C
\
BY:
$
(LSI-FIB-UPC)
Notes
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
4 / 20
Introduction
Systems that think like humans
Notes
The model is how the human mind works
By experimentation we obtain a theory about how the mind works
(psychological experiments)
From this theory we can build computational models
Cognitive science
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
5 / 20
Introduction
Systems that think rationally
Notes
The laws of rationality are based on logic
Formal logic has to be the basis of intelligent systems (Logicism)
There are two obstacles:
It is very difficult to formalize knowledge in logic language
There is a great leap between the theoretical capacity of logic and its
practical application
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
6 / 20
Introduction to Artificial Intelligence
Introduction
Systems that act rationally
Notes
To act rationally means to perform acts to achieve the best outcome,
to obtain some goals given some beliefs
The paradigm is the agent
An agent perceives and act, accordingly to the environment it is
situated
The capacities that are needed are the same than those to pass the
Turing’s test: Natural language processing, knowledge representation,
reasoning, learning, perception
It is a more general approach, not centered on the human model
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
7 / 20
Introduction
Is Artificial Intelligence possible?
Notes
The possibility to build an artificial intelligence raises some complex
philosophical problems
Are thinking machines conscious?
The chinese room, (Searle, 1980)
Is intelligence an emerging property of the biological elements that
sustains it?
There is no definitive conclusion
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
8 / 20
Introduction
The chinese room
C
\
BY:
$
(LSI-FIB-UPC)
Notes
+
−>
+
−>
+
−>
Artificial Intelligence
Term 2009/2010
9 / 20
Introduction to Artificial Intelligence
Foundations of Artificial Intelligence
Foundations of Artificial Intelligence - Philosophy
Notes
Are mechanical intelligences possible?
Aristotle
Llull
Decart
Laws that govern the rational part
Actions are justified by a connection bet-
Knowledge = Logic Theories
Aristotle
C
\
Perception is the source of knowledge
(Induction)
ween goals and knowledge
BY:
$
(LSI-FIB-UPC)
Leibniz
The mind is a physical system
of the mind (logic)
Russell
Artificial Intelligence
Introduction to Artificial Intelligence
Hume
Bacon
Term 2009/2010
10 / 20
Foundations of Artificial Intelligence
Foundations of Artificial Intelligence - Mathematics
Notes
Mathematics
Logic
Boole
Complexity
Frege
What are the rules of
reasoning?
C
\
BY:
$
(LSI-FIB-UPC)
Gödel
Turing
What is computable?
Probability
Fermat Bernoulli Bayes
How to reason with uncertainty?
Artificial Intelligence
Introduction to Artificial Intelligence
Term 2009/2010
11 / 20
Foundations of Artificial Intelligence
Foundations of Artificial Intelligence (III)
Notes
Economics
¿How do we make decisions ...
that benefit us?
against competitors?
when there is not immediate benefit?
Decision theory/Game theory/Operations research
Neuroscience
How does the brain process information?
Neurons/Specialized areas in the brain
Psychology
How does people act and think?
Cognitive psychology/Cognitive science: Behavioural theories,
foundations of rational behaviour
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
12 / 20
Introduction to Artificial Intelligence
Foundations of Artificial Intelligence
Foundations of Artificial Intelligence (IV)
Notes
Computer Science
In order AI to exist it is needed a physical device to support it
(Hardware)
AI needs also software tools to develop intelligent systems
Control theory/Cybernetics
Development of autonomous systems
Linguistics
Chomsky: Knowledge representation, language grammar
Computational linguistics
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
13 / 20
Areas of Artificial Intelligence
Areas of Artificial Intelligence
Notes
Basic areas
Knowledge representation
Automatic problem solving, heuristic search
Specific areas
Automatic planning
Natural language processing
Automatic reasoning
Knowledge based systems
Perception
Machine learning
Intelligent agents
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Areas of Artificial Intelligence
Term 2009/2010
14 / 20
Applications of Artificial Intelligence
Autonomous robots
Notes
Autonomous Navigation
Assistive Technologies
Complex tasks (vision, planning, coordination, real time, ...)
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
15 / 20
Areas of Artificial Intelligence
Applications of Artificial Intelligence
Image recognition
Faces
Medical Images
Hand writing recognition
Object recognition
BY:
$
(LSI-FIB-UPC)
C
\
Notes
Artificial Intelligence
Areas of Artificial Intelligence
Term 2009/2010
16 / 20
Applications of Artificial Intelligence
Computer games
Notes
Path finding
Strategy
Coordination, cooperation, learning, adaptation, ...
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Areas of Artificial Intelligence
Term 2009/2010
17 / 20
Applications of Artificial Intelligence
Electronics appliances
Intelligent vacuum cleaners
Notes
Cameras with face recognition
Apliances with intelligent control
C
\
BY:
$
(LSI-FIB-UPC)
Artificial Intelligence
Term 2009/2010
18 / 20
Areas of Artificial Intelligence
Applications of Artificial Intelligence
Intelligent Interfaces/Recommendation/Personalization
Ambient Intelligence
C
\
BY:
$
(LSI-FIB-UPC)
Notes
Recommendation/Personalization
Artificial Intelligence
Areas of Artificial Intelligence
Term 2009/2010
19 / 20
Applications of Artificial Intelligence
Diagnostic/Control/Design/Planning systems
Notes
Medicine
Logistics
Intelligent Manufacturing
Industrials Processes
Industrial Design
Complex Processes
(LSI-FIB-UPC)
\
$
C
BY:
Artificial Intelligence
Term 2009/2010
20 / 20
Notes