X-Message-Number: 15183 From: "Gary Tripp" <> Subject: real-time simulation issues Date: Mon, 25 Dec 2000 11:37:08 -0500 This is a multi-part message in MIME format. ------=_NextPart_000_000D_01C06E67.0330F1C0 Content-Type: text/plain; charset="iso-8859-1" Thomas is concerned that the fact that the real time interaction of our neural structures with themselves and the environment may not be adequately simulated by a computer. We usually think of computers in terms of a Von-Neuman architecture whereas it may be possible to fashion super massively parallel switching systems with elements that operate perhaps 10's of millions of times faster than neurons. In theory we could even go so far as to construct neural nets employing very fast switches. I expect that at some point in the future we will be able to dispatch nanoprobes to strategic locations in our biological neural nets and instruct them to monitor the behaviour of real neurons so that we can learn how they function in a net. With this information we could build perfect artificial neural nets. /gary ------=_NextPart_000_000D_01C06E67.0330F1C0 Content-Type: text/html; [ AUTOMATICALLY SKIPPING HTML ENCODING! ] Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=15183