In your recent "How is sensory information processed?", q-bio.NC/0402029, you wrote:

"Advances in statistical learning theory leave us with many possible designs of learning machines. But which of them are implemented by brains, metabolic and genetic networks, and other biological information processors?
...
"Neuronal and neural networks are usually drawn as homogeneous, layered, and often self-similar objects."

I'm afraid this is not true. See some textbook in neurohistology.

"They are also believed to be fairly general computational machines, able to solve the most diverse problems."

Nope. They can't solve even a simple self-referential paradox (the halting problem in Turing machines). We can. See

http://members.aon.at/chakalov/ReflexiveParadoxes.html

http://members.aon.at/chakalov/Penrose.html#NB

I suppose you and Chris Wiggins met Willian Bialek in Princeton during your Ph.D. programs. See his "Thinking about the brain", physics/0205030,

http://members.aon.at/chakalov/Adler.html#2

and recall that you can think *about* your brain, *with* your brain,

http://members.aon.at/chakalov/Beauregard.html

No machine can perform this self-acting.

I suppose you've been brainwashed by Marxist-Leninist philosophy (I believe you're still fluent in Russian), but I cannot understand your colleague W. Bialek. Is he Russian?

Sincerely,

Dimi Chakalov
http://members.aon.at/chakalov
http://members.aon.at/chakalov/white_paper.html
--
Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be at the rghit pclae.  The rset can be a total mses and you can sitll raed it wouthit a porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe,

http://members.aon.at/chakalov/faq.html

Pritie amzanig huh?