From newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!ames!agate!doc.ic.ac.uk!uknet!mcsun!sunic!dkuug!imada!breese Wed Sep 23 16:54:11 EDT 1992
Article 6955 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!ames!agate!doc.ic.ac.uk!uknet!mcsun!sunic!dkuug!imada!breese
>From: breese@imada.ou.dk (Bjoern Reese)
Newsgroups: comp.ai.philosophy
Subject: Empty memory spots?
Message-ID: <1992Sep15.113556.7474@imada.ou.dk>
Date: 15 Sep 92 11:35:56 GMT
Sender: news@imada.ou.dk (USENET News System)
Organization: Dept. of Math. & Computer Science, Odense University, Denmark
Lines: 35


Are there empty spots in the memory?
------------------------------------

On rare occations a flash of inspiration hits me. So too, the other day. I've
always considered the mind to be full of empty, unused areas, where new input
could be stored. I have lately begun to question this view.

My new belief is that all the memory is used all the time. All memory is filled
up with mirrored and hence superfluous data. New data is stored on top of some
other data (hopefully stored elsewhere in our holographic memory). This means,
whenever you learn something new, your memory of things you already know is
becomming more and more weak. Also, the data are scattered throughout the
memory, which could be a reason for our irrational thoughts popping up every
now and then.

Let's take a newborn child, who only knows two things (the rest is considered
reflex, for the sake of simplicity): 1) the warm and cosy embryonic life
before birth, and 2) the cold and unstable world afterwards. Its memory is
totally filled up with these two experiences, and therefore its very intense
(I cried a lot too :-). As it learns more and more of the world around it, the
intensity of the original remembrance is gradually fading.

The amount of `backups' is decreasing as the amount of non-redundancy data
increases.

Perhaps someone more clear-headed or more clear-sighted can shed some light
into my shaded mind? Perhaps it could hold an interest for the people of
neural networks?

- breese


PS: If this post has already appeared, please forgive me. I'm learning how
    to use `nn' right now (and got Distribution wrong several times).


