| | |
Summary: 1
Intelligence, Sex and the Single Neuron a Position Paper
Paul Adams and Kingsley Cox SUNY, Stony Brook, NY 11790
Some New Principles Our unconventional approach to understanding human
intelligence is rooted in neuroscience, specifically the assumption that to make progress
one must first understand the basic principles underlying neocortical circuitry, which is
shared by all mammals. Thus, rather than frontal attack, we advocate siege. Our second
unusual premise is that the only well-understood case of biological intelligence,
Darwinian evolution, could provide essential clues. By this we do not mean that a
combination of genetic algorithms and neural networks could underlie human
intelligence, but instead that the underlying logic of these 2 forms of intelligence
(indeed, of any form) might be similar. One important clue is that "biology" falls into 2
parts: our understanding of the general machinery common to all life (genes/DNA,
proteins, cells etc) and the myriad, concrete, apparently intelligent, solutions to specific
problems that different species evince. One could never hope to understand the former
by studying the details of the latter, though this is the strategy used by almost all those
interested in neural intelligence. But the most important clue is that superaccurate
replication is essential for evolution and "Darwinian intelligent design". Thirdly, we
assume, conventionally, that learning is key to intelligence. But since the brain is
composed of neurons, we think, rather surprisingly, that "intelligence" can emerge even
|
