Cast as the balance between internal representation and external regularities, learning has long been a central issue in the understanding of intelligence. It may be argued however that until recently its indispensable role had not been entirely embraced in computer science. Indeed, the deenition of intelligence that is frequently invoked in the eld of artiicial intelligence does not quite provide learning its deserved stature. For many years, the Turing test has served as an operational deenition of intelligence, using which, most investigators in artiicial intelligence have either directly or implicitly measured their goals and achievements: If a computer behaves in a way indistinguishable from a human, then it can be called intelligent, even if this \intelligence" has been painstakingly programmed by a highly skilled individual. Consistent with Turing's implicit deenition, intelligence was perceived thirty years ago as mainly reasoning, problem solving, theorem proving, or game playing (chess, for instance). In contrast, today, we understand intelligent lower animals better and comprehend the complexities of the tasks that our senses routinely solve. We have also realized through experience how intractable the problem of producing software can prove, and how much of it would be needed to replicate even the simplest aspects of intelligence. In light of these submissions, it may be argued that a somewhat altered deenition of intelligence may be more suited to our present notions of the underlying problems. It seems to us that such a modiied Turing test should emphasize the development and learning of perceptual, motor, and language competence. In particular, many would agree that a system should be declared intelligent only if it is capable of rst acquiring and then exhibiting motor, perceptual, and language skills. This revised Turing test expresses, in an explicit manner, the central role of learning in the art and science of replicating intelligence. That the brain is not hard-wired but rather can be modiied by experience 1
[1]
Pavel Pudil,et al.
Introduction to Statistical Pattern Recognition
,
2006
.
[2]
Editors
,
2003
.
[3]
Pat Langley,et al.
Elements of Machine Learning
,
1995
.
[4]
Azriel Rosenfeld,et al.
Machine Vision and Learning: Research Issues and Directions
,
1994
.
[5]
S. Hyakin,et al.
Neural Networks: A Comprehensive Foundation
,
1994
.
[6]
M. Alexander,et al.
Principles of Neural Science
,
1981
.
[7]
Richard O. Duda,et al.
Pattern classification and scene analysis
,
1974,
A Wiley-Interscience publication.