Biomedical Engineering Reference
In-Depth Information
Due to the ability of the human mind to quickly
grasp the recipes of a problem's solution that an
artificial system has applied, the question remains
open whether solutions found by a GP system
will remain qualitatively better than solutions
discovered by humans over the long term. Per-
haps the best area to consider for this kind of
attempt is mathematics. First results have been
achieved that seem to indicate that, under very
special circumstances, certain mathematical prob-
lems can be solved more efficiently using GP
[98]
.
machines is still far in the future; however, in
restricted steps in that direction have been taken.
It is the firm conviction of the author of this
chapter that a major component of any future
system that could truly lay claim to the property
of intelligence will be bioinspiration.
Acknowledgments
I express my sincere gratitude to my students, collabora-
tors, and colleagues with whom it is such a pleasure to work
on various projects in evolutionary computation. I also
acknowledge funding agencies that financed many proj-
ects over the course of nearly two decades. Specifically, I
mention the German Science Foundation (DFG), the Technical
University of Dortmund, and the state of Northrhine-
Westphalia for funding from 1993 to 2003 and NSERC (Canada),
Memorial University of Newfoundland, and the government
of Newfoundland for funding from 2003 to present.
17.7 CONCLUSIONS
Implementation of GP will continue to ben-
efit in coming years from new approaches that
include results from developmental biology and
epigenetics.
Application of GP will continue to broaden.
Many applications focus on engineering appli-
cations. In this role, GP may contribute consid-
erably to creative solutions to long-held
problems in the real world.
Since GP was first used around 1990, raw
computational power has increased by roughly
a factor of 40,000 following Moore's law of dou-
bling of transistor density every 18 months. As
Koza points out, although initially only toy
problems were amenable to solution through
GP, subsequent increases in computational
power and methodological progress of GP have
allowed new solutions to previously patented
inventions as well as, more recently, completely
new inventions that are by themselves patent-
able. A milestone in this regard was reached in
2005 when the first patent was issued for an
invention produced by a GP system
[99]
.
The use of bioinspiration, notably through
lessons from our understanding of natural evo-
lution, has led to some very substantial progress
in the implementation of artificial systems that
show human-level problem-solving abilities.
Achieving
artiicial intelligence
in computing
References
[1] C. Darwin,
On the origin of species
, John Murray, London,
UK (1859).
[2] C.M. Bishop,
Neural networks for pattern recognition
,
Clarendon Press, Oxford, UK (1995).
[3] G.J. Klir and B. Yuan,
Fuzzy sets and fuzzy logic
, Prentice-
Hall, Englewood Cliffs, NJ, USA (1995).
[4] L.N. De Castro and J. Timmis,
Artificial immune systems:
a new computational intelligence approach
, Springer-Verlag,
Heidelberg, Germany (2002).
[5] M. Dorigo and C. Blum, Ant colony optimization
theory: a survey,
Theor Comput Sci
344
(2005), 243-278.
[6] E. Bonabeau, M. Dorigo, and G. Theraulaz,
Swarm intel-
ligence: from natural to artificial systems
, Oxford Univer-
sity Press, New York, NY, USA (1999).
[7] C. Adami,
Introduction to artificial life
, Telos, Springer-
Verlag, New York, NY, USA (1998).
[8] A.M. Turing, Computing machinery and intelligence,
Mind
59
(1950), 433-460.
[9] J.H. Holland, Outline for a logical theory of adaptive
systems,
J ACM
9
(1962), 297-314.
[10] J. Holland,
Adaptation in natural and artificial systems
,
University of Michigan Press, Ann Arbor, MI, USA
(1975).
[11] R.M. Friedberg, A learning machine: Part I,
IBM J Res
Dev
2
(1958), 2-13.
[12] L. Fogel, A. Owens, and M. Walsh,
Artificial intelligence
through simulated evolution
, Wiley, New York, NY, USA
(1966).
