X-Message-Number: 4314
From: 
Date: Sat, 29 Apr 1995 00:27:11 -0700
Subject: Neural nets

In message #4192, Keith F. Lynch asserts:

>> Neural nets are computationally equivalent to conventional 
>> computers.

In message #4232, John Clark asserts:

>> but any problem that can be solved by a neural net can be
>> solved by a Turing machine;  mathematically and 
>> philosophically the two are equivalent."

In message #4308 Eugen Leitl makes the same assertion.  In 
message #4236, Paul Wakfer makes the same assertion.

All of these persons are claiming that every recursive function 
can be computed by a suitable neural net. I have asked both Lynch 
and Clark, in private e-mail, if they can cite a reference to the 
literature wherein this claim is proved. Neither has yet 
responded.

So I ask:  Can anyone who believes this claim cite a reference to 
the literature wherein it is demonstrated that every recursive 
function can be computed by a suitable neural net?


Art Quaife
Trans Time, Inc.


Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=4314