Q: Implementation of an Associative Memory Lookup Index


Subject: Q: Implementation of an Associative Memory Lookup Index
From: Someone Insignificant (mike@vtracks.com)
Date: Sat Feb 19 2000 - 01:47:44 MET


Hi All!

Ive been scouring the net for days looking for a solution to a particular
problem, and my research has led me here to the AI group.

I have a database application, implemented in C on a Linux platform, that
needs to keep track of { document-id }{ location-id } pairs, approximately
300 million of them on a single machine. The best solution ( ie fast
lookups, minimal memory usage ) that I have been
able to implement this far is a modified ternary trie structure, however,
the trie structure utilizes far to much memory ( each id is 32-bits ).

After searching through reams of documents and reasearch papers, I found a
company that is doing this using simulated
content addressable memory. However, they are currently charging $25,000 a
copy for their software and thats about $25,000 over
what my budget allows. In case your interested, the product is called
NeoData and you can find a white paper on it at http://www.neocore.com/

Im not an AI expert by any means, but after doing some more research, Ive
ended up here looking for such a solution to my problem and
from what I understand, a Hopefield Network might be able to do exactly
this.

Given a 32-bit document Id, the network would produce a location id
corresponding to the location of the document.

Anyway, am I way off the mark here? Will a Hopefield Network work for this
problem? Am I completely clueless? And, if so, can you point me in the right
direction?

Thanks in advance and good luck giving birth to any digital babies that you
might be brewing.

mike
mike@vtracks.com

[ comp.ai is moderated. To submit, just post and be patient, or if ]
[ that fails mail your article to <comp-ai@moderators.isc.org>, and ]
[ ask your news administrator to fix the problems with your system. ]

############################################################################
This message was posted through the fuzzy mailing list.
(1) To subscribe to this mailing list, send a message body of
"SUB FUZZY-MAIL myFirstName mySurname" to listproc@dbai.tuwien.ac.at
(2) To unsubscribe from this mailing list, send a message body of
"UNSUB FUZZY-MAIL" or "UNSUB FUZZY-MAIL yoursubscription@email.address.com"
to listproc@dbai.tuwien.ac.at
(3) To reach the human who maintains the list, send mail to
fuzzy-owner@dbai.tuwien.ac.at
(4) WWW access and other information on Fuzzy Sets and Logic see
http://www.dbai.tuwien.ac.at/ftp/mlowner/fuzzy-mail.info
(5) WWW archive: http://www.dbai.tuwien.ac.at/marchives/fuzzy-mail/index.html



This archive was generated by hypermail 2b25 : Thu Apr 06 2000 - 15:59:45 MET DST