Newsgroups: comp.ai.nat-lang,comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!das-news.harvard.edu!news2.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!pipex!uknet!festival!edcogsci!kversp
From: kversp@cogsci.ed.ac.uk (Karin Verspoor)
Subject: Re: PAPER ANNOUNCEMENT (UPDATE)
Message-ID: <CvnI1C.6L5@cogsci.ed.ac.uk>
Keywords: natural language neural networks machine translation
Organization: Centre for Cognitive Science, Edinburgh, UK
References: <nk3.778530751@dolphin.doc.ic.ac.uk> <3482ja$4k9@venera.isi.edu>
Date: Mon, 5 Sep 1994 09:51:10 GMT
Lines: 26
Xref: glinda.oz.cs.cmu.edu comp.ai.nat-lang:2035 comp.ai.neural-nets:18723

In article <3482ja$4k9@venera.isi.edu> haines@isi.edu (Matthew Haines) writes:
>
>Before you go running off to FTP the paper, beware that 
>Koncar's system uses only 64 words per language, the input
>sentences were limited to eight words, and the
>output sentences six.  Furthermore, the grammar was 
>limited to "a simple deterministic grammar."
>
>If I've ever seen a toy system, this is it.

Ummm, excuse me, but shouldn't the criteria be *quality* not
*quantity*?

As with every new endeavour, you have to start small and work your way
up...

Just my $.02.

Karin Verspoo


-- 
----------------------------------------------------------
Karin Verspoor                                 _\\|//_
MSc student                                      o o
Centre for Cognitive Science               --oo0--U--0oo--
