Get lost, Mr. Cochlea! ("John K. Bates" )


Subject: Get lost, Mr. Cochlea!
From:    "John K. Bates"  <jkbates(at)COMPUTER.NET>
Date:    Tue, 27 Feb 2001 10:38:36 -0500

Dear Yadong, I applaud your graphic dialog between brain and cochlea. What this discussion does is to expose further the need for an operational system anlysis that predicts how the brain could guide survival based on fundamental requirements and constraints. The object is to rationalize how waveshape information could be related to the meaning that is needed to make the survival decisions. The basic error in current thinking is, in my view, in trying to find a purely mathematical basis (a representation or a code) that applies to understanding the process of selecting what is important in a signal ensemble. The existence of mathematical "information" does not directly relate to meaning. And meaning is what determines survival, which is what Mr. Brain's job is. In carrying out this process, there appears to be no deterministic way to "calculate" meaning. And meaning is what Mr. Brain uses to decide at at any point in time which part of Mr. Cochlea's information is useful and which part to ignore. (I said ignore, not discard, because contrary to Ellis's view, we're speaking of attention and awareness. Mr. Brain can't survive without having awareness simultaneously with attention.) The task for CASA, then, is to implement a computational algorithm that can somehow represent the perception of meaning. I happen to believe that such an algorithm is attainable, at least in rudimentary or specialized form. (For example, the quantifying of human emotion is being explored in bimodal speech analysis, as described in the current issue of the Signal Processing Magazine.) But finding the correct approach must begin by responding to Seebeck's admonition to Ohm and von Helmholtz to understand what the ear hears before imposing an arbitrary theory on the process. By the way, I have added to my Web site a paper describing how my zero-based particulate algorithm might be used in a cochlear implant. It might shed some light on the existence of a suitable temporal cochlear representation. -John Bates Time/Space Systems Pleasantville, NY 10570 <http://home.computer.net/~jkbates>


This message came from the mail archive
http://www.auditory.org/postings/2001/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University