Graphics Reference
In-Depth Information
Artificial Intelligence
Creating or simulating intelligence require
several computer's capabilities (Luger, 2004)
including reasoning, perception (with object and
facial recognition), deduction, problem solving,
creativity and imagination, general and social
intelligence and skills, knowledge representation
(including ontologies representing concepts and
relationships), planning of actions (including
multi-agent planning, emergent behavior as dis-
played in evolutionary algorithms and in swarm
intelligence) machine learning, natural language
processing (including text mining for information
retrieval and machine translation), robotic motion
and object manipulation, among other problems.
Scientists working on biologically inspired
artificial intelligence (Floreano & Mattiussi,
2008) are studying theories and observations of
biological systems that result from evolutionary
processes: populations' diversity, heredity, and
selection processes leading to genetic evolution
and mutations. They develop methods to combine
engineering with technologies aimed at construct-
ing evolutionary algorithms to create collective
systems presenting self-organization, particle
swarm organization, ant colony organization, and
swarm robotics, along with developing models of
evolutionary dynamics, competition, and coopera-
tion. Artificial systems characterized by the bio-
inspired artificial intelligence include evolutionary
systems and artificial evolutionary developmental
systems, cellular systems and cellular automata,
artificial and hybrid neural networks, artificial
immune systems, and behavioral systems, among
other approaches.
Computer scientists construct intelligent systems
able to process information in an intelligent way.
This part of computer science, which began about
1956 when the computer scientist John McCarthy
coined this term, focuses on the study and design
of intelligent agents, which perceive their environ-
ment and take the most possible successful ac-
tions. Tools for accomplishing these goals include
mathematically based search and optimization,
logic, probability methods based on computing
for economics, and others.
Advancing the artificial intelligence domain
involves creating agents that think humanly and
rationally as well as act humanly and rationally
(Russel & Norvig, 2009). Systems of agents
and their societies may include software, robots,
and humans. Multi agent systems (MAS) may
interact with each other and/or with an environ-
ment to exchange data and support human social
activities such as communication, cooperation,
coordination, negotiation, sharing, or competition.
Intelligent agents may observe through sensors
to act rationally (learn and use knowledge) by
taking autonomous actions aimed at performing
their design objectives (Trajkovski, 2009, 2010).
As stated by Boden (2006),
Cognitive science uses abstract (logical/math-
ematical) concepts drawn from artificial intel-
ligence (AI) and control theory, alias cybernetics.
•
AI tries to make computers do the sorts
of things that minds can do. These things
range from interpreting language or cam-
era input, through making medical diagno-
ses, and constructing imaginary (virtual)
worlds, to controlling the movements of a
robot.
Singularity
The writer Vernor Vinge popularized the term
singularity; he believed that artificial intelligence,
human biological enhancement, and computer
cognitive abilities would surpass these of any
human being. Some researchers hypothesize the
future emergence of greater-than-human super
intelligence, partly because people would undergo
•
Control theory studies the functioning of
self-regulating systems. These systems in-
clude both automated chemical factories
and living cells and organisms. (Boden,
2006, p. 4)
