Newsgroups: comp.robotics
Path: brunix!cat.cis.Brown.EDU!agate!howland.reston.ans.net!math.ohio-state.edu!jussieu.fr!univ-lyon1.fr!swidir.switch.ch!newsfeed.ACO.net!Austria.EU.net!EU.net!uunet!hobbes!earth.armory.com!rstevew
From: rstevew@armory.com (Richard Steven Walz)
Subject: Re: 100 Billion Nuerons
Organization: The Armory
Date: Tue, 19 Jul 1994 18:00:57 GMT
Message-ID: <Ct78pn.L1@armory.com>
References: <PgyePc1w165w@sfrsa.com> <4ZoJPc1w165w@sfrsa.com>
Sender: news@armory.com (Usenet News)
Nntp-Posting-Host: deeptht.armory.com
Lines: 28

In article <4ZoJPc1w165w@sfrsa.com>, bsmall <bsmall@sfrsa.com> wrote:
>eds@sfrsa.com (Ed Severinghaus) writes:
>
>> 
>> I am not supporting the counter notion, but in terms of "touching" every 
>> cell regularly, this could be made semi-automatic ("low-level").  Also 
>> consider the process in a dynamic RAM chip!!!  
>> 
>> Dynamically yours,
>>  
>> Ed
>> 
>
> 
>My gut reaction is that dynamic RAM decay is too fast. I'd
>also not want to get intermittent problems because certain 
>bits were decaying faster than others. Is it true that RAM
>decays to 0's?
> 
>Another problem that I was thinking about was fatigue. If 
>subjected to a long string of stimulus, a neuron fatigues.
>I believe this is an important part of the thinking process.
>Will this intail yet more counters?
------------------------------------
When I had a refresh failure in one row of DRAM once, it all went to FFH,
or all ones, so I think it decays to ones. Anyone?
-Steve Walz

