Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!oitnews.harvard.edu!yale!yale.edu!spool.mu.edu!howland.reston.ans.net!vixen.cso.uiuc.edu!uchinews!kimbark!gal2
From: gal2@kimbark.uchicago.edu (Jacob Galley)
Subject: Re: ART2 or ART2-a source code needed
X-Nntp-Posting-Host: midway.uchicago.edu
Message-ID: <DFoCyz.EH7@midway.uchicago.edu>
Sender: news@midway.uchicago.edu (News Administrator)
Reply-To: gal2@midway.uchicago.edu
Organization: University of Chicago
References: <joakim.waldemark.4.000B05A1@teknikum.umu.se> <43qfnk$2fr@news.magi.com> <95264.165523JXDGC@CUNYVM.CUNY.EDU> <444hje$iba@st-james.comp.vuw.ac.nz>
Date: Fri, 29 Sep 1995 16:11:23 GMT
Lines: 31

Guszti.Bartfai@comp.vuw.ac.nz (Guszti Bartfai) writes:
>
><JXDGC@CUNYVM.CUNY.EDU> writes:
>|>
>|> Could someone please post references to ARTICLES about ART,
>|> as opposed to source code that implements it. I have been hearing a lot
>|> about them, but haven't been able to READ about them.
>
>Here are some ARTicles:

	<snip>

>Carpenter and Ross, 1993
>   Carpenter, G. and Ross, W. (1993). ART-EMAP: A neural network architecture
>   for learning and prediction by evidence accumulation. Technical Report
>   CNS/CAS-TR-93-015, Boston University, Boston, MA.

This article has been published in _IEEE Transactions on Neural Networks_, 
July 1995, 6(4):805-818.

It basically describes four independent but complimentary
augmentations to Fuzzy ARTMAP.  The idea is that if an input does not
match any learned category very well, the system should hold onto the
information, make another "observation" of the same stimulus, and
commit to a category based on the accumulation of evidence through
repeated observations.  Interesting!

Jake.

-- 
Reinheitsgebot <-- "Keep your laws off my beer!" <-- gal2@midway.uchicago.edu
