Biomedical Engineering Reference
In-Depth Information
patterns of cells seeking stable repeating patterns, or patterns that move like the gliders. An
interesting aspect of this game was that the patterns found by computers were discovered rather
than invented.
Some of the benefits of using computers have been the development of the ''genetic program-
ming'' or ''evolutionary programming'' (Chapters 4 and 5; Koza, 1992). The ''DNA'' of genetic
programming consists of a set of equations and operations where the computer software measures
how well each program solves a particular problem. The programs that fare the worst are eliminated
and new strains of program code are bred by recombination, either with or without mutation. The
solutions produced by evolutionary programming emulate the solutions in the real world, and it
may use functions that seemingly have no logical relevance to the problem that is being solved but
it produces effective solutions (Chapters 4 and 5).
1.4 ARTIFICIAL INTELLIGENCE
According to the American Association for Artificial Intelligence (AAAI), artificial intelligence
(AI) is, ''the scientific understanding of the mechanisms underlying thought and intelligent behav-
ior and their embodiment in machines.'' AI is a branch of computer science that studies the
computational requirements for such tasks as perception, reasoning, and learning, to allow the
development of systems that perform these capabilities (Russell and Norvig, 2003). AI researchers
are addressing a wide range of problems that include studying the requirements for expert
performance of specialized tasks, explaining behaviors in terms of low-level processes, using
models inspired by the computation of the brain, and explaining them in terms of higher-level
psychological constructs such as plans and goals. The field seeks to advance the understanding of
human cognition (Chapter 3), understand the requirements for intelligence in general, and develop
artifacts such as intelligent devices, autonomous agents, and systems that cooperate with humans to
enhance their abilities. The name AI was coined in 1956, though the roots of the field may be
attributed to the efforts in World War II to crack enemy codes by capturing human intelligence in a
machine that was called Enigma. This approach eventually led to the 1997 computer success of
IBM's Deep Blue in beating the world-champion chess player Garry Kasparov. Even though this
was an enormous success for computers, it still does not resemble human intelligence. AI tech-
nologies consist of an increasing number of tools, including artificial neural networks, expert
systems, fuzzy logic, and genetic algorithms (Luger, 2001; Chapters 4 and 5).
Advances in AI are allowing analysis of complex nonlinear problems that are beyond the
capability of conventional methods by using such tools as neural networks (i.e., networks of
artificial brain cells) that can learn and recognize patterns and reach solutions. This is providing
enormous capabilities in the area of robotics including the ability to operate autonomously. One of
the milestones in AI is the development of ''Shakey'' robot, which was completed by SRI
International's Artificial Intelligence Center (AIC) in 1972. This six-foot tall robot (http://www-
clmc.usc.edu/~cs545/Lecture_I.pdf) was named for its erratic and jerky movement. Shakey is the
first mobile robot to visually interpret its environment, locate items, navigate around them, and
reason about its actions. Shakey was equipped with a TV camera, a triangulating range finder,
bumpers, and a wireless video system and it has the capability of autonomous decision making.
The subject of AI is widely covered in the literature (e.g., Luger, 2001; Russell and
Norvig, 2003). Chapter 3 of this topic addresses the topic of modeling computers after the processes
in the human brain. One area of AI, which mimics nature, is the swarm intelligence that involves
the study of self-organizing processes in artifacts of nature and humans. Algorithms inspired by
social insect behavior have been proposed to solve difficult computational problems such
as discrete optimization where the ant colony optimization process was followed. Resulting
algorithms were used to solve such problems as vehicle routing and routing in telecommunication
networks.
Search WWH ::
Custom Search