Steven Sloman wrote:
- -----Original Message-----
From: Steven Sloman [mailto:[EMAIL PROTECTED]
Sent: Thursday, November 20, 2003 10:08 AM
To: [EMAIL PROTECTED]
Subject: Re: [UAI] Useful algorithms based on human inference
<snip>
By the way, the "neural nets" that inspired the connectionist
revolution of the mid-1980s were not inspired primarily by the brain
but rather by cognitive phenomena. The best-known book of the time,
Rumelhart and McClelland's volumes on Parallel Distributed Processing,
has very little to say about the brain but includes many important
models of cognitive phenomena involving, for example, memory and
classification.
steven
Steven Sloman
Cognitive & Linguistic Sciences
Brown University, Box 1978
Providence, RI 02912
email: [EMAIL PROTECTED]
http://www.cog.brown.edu/~sloman
========================
I beg to differ. Neural nets date back to the idealized neurons of Pitts and
McCullough in the early 40s and most of the early work in the 40s, 50s, and 60s was
explicitly aimed at modeling the behavior of the central nervous system and exploring
the stability and emergent properties of networks in general (e.g., Ashby's homeostat
and distributed models -- that is, holographic -- of memory). There was also an early
thread of research aimed at using the best understanding of how the nervous system did
pattern recognition, for instance, to develop electrical/electronic systems that could
do the same, leading eventually to Rumelhart and McClelland, instrumental in the
revival of interest in neural networks, but building on four decades of prior research.
Best regards.
Jim Dukelow
Pacific Northwest National Laboratory
Richland, WA
[EMAIL PROTECTED]
These comments are mine and have not been reviewed and/or approved by my management or
by the U.S. Department of Energy.
