Send Moses-support mailing list submissions to
moses-support@mit.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.mit.edu/mailman/listinfo/moses-support
or, via email, send a message with subject or body 'help' to
moses-support-request@mit.edu
You can reach the person managing the list at
moses-support-owner@mit.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Moses-support digest..."
Today's Topics:
1. Learning curves (J?rg Tiedemann)
2. Re: Learning curves (marco turchi)
3. Re: Learning curves (Prasanth K)
4. normalization issue in tokenization of Kannada words in
baseline MT (Anoop (?????))
----------------------------------------------------------------------
Message: 1
Date: Fri, 21 Nov 2014 17:18:07 +0100
From: J?rg Tiedemann <tiedeman@gmail.com>
Subject: [Moses-support] Learning curves
To: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID: <5AA19DAC-6DEB-440F-AD14-3D5394BEFB8E@gmail.com>
Content-Type: text/plain; charset=utf-8
Hi,
Does anyone have good pictures that show the correlation between data size and Bleu scores? Both, LM and TM training would be interesting. Thanks!
J?rg
------------------------------
Message: 2
Date: Fri, 21 Nov 2014 17:37:33 +0100
From: marco turchi <marco.turchi@gmail.com>
Subject: Re: [Moses-support] Learning curves
To: J?rg Tiedemann <tiedeman@gmail.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CABEbM3-uM0ywamLBNKZ0Z3bXg90kF3t+-=NknXOOG5P3rFuhdw@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi Jorg,
I published these papers on this topic few years ago. I hope they can be
useful.
"Learning performance of a machine translation system: a statistical and
computational analysis." *Proceedings of the Third Workshop on Statistical
Machine Translation*. Association for Computational Linguistics, 2008.
"Learning to translate: a statistical and computational analysis." *Advances
in Artificial Intelligence* 2012 (2012): 1.
"Learning Machine Translation from In-domain and Out-of-domain Data."
*Proceedings
of the 16th Annual Conference of the European Association for Machine
Translation*. 2012.
Thanks a lot
Marco
On Fri, Nov 21, 2014 at 5:18 PM, J?rg Tiedemann <tiedeman@gmail.com> wrote:
> Hi,
>
> Does anyone have good pictures that show the correlation between data size
> and Bleu scores? Both, LM and TM training would be interesting. Thanks!
>
> J?rg
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20141121/c83280fc/attachment-0001.htm
------------------------------
Message: 3
Date: Fri, 21 Nov 2014 17:39:31 +0100
From: Prasanth K <prasanthk.ms09@gmail.com>
Subject: Re: [Moses-support] Learning curves
To: J?rg Tiedemann <tiedeman@gmail.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CA+n+9-ghgZek58JcgGNrgV37SMpjSfGrcZDVbDPDQ=ap02CVLg@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi Jorg,
I recall Marco Tuchi had some statistical analysis about these under
different conditions for translating from Spanish-English.
I did have to create "lots of good pictures" for my own paper for different
language pairs and corpora. And the data i.e. BLEU score for different
training data sizes can be found in the supplementary data along with the
paper. Hope that should be helpful.
http://aclweb.org/anthology/P/P12/P12-1003.pdf
- Regards,
Prasanth
On Fri, Nov 21, 2014 at 5:18 PM, J?rg Tiedemann <tiedeman@gmail.com> wrote:
> Hi,
>
> Does anyone have good pictures that show the correlation between data size
> and Bleu scores? Both, LM and TM training would be interesting. Thanks!
>
> J?rg
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
--
"Theories have four stages of acceptance. i) this is worthless nonsense;
ii) this is an interesting, but perverse, point of view, iii) this is true,
but quite unimportant; iv) I always said so."
--- J.B.S. Haldane
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20141121/13439bdd/attachment-0001.htm
------------------------------
Message: 4
Date: Fri, 21 Nov 2014 22:43:42 +0530
From: Anoop (?????) <anoop.kunchukuttan@gmail.com>
Subject: [Moses-support] normalization issue in tokenization of
Kannada words in baseline MT
To: moses-support@mit.edu
Message-ID:
<CADXxMYeTsvU_48vCS_nDqR5qKDe-D47fPuUVz7-T8d=H=Ri7Mg@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"
Hi Shiva,
The Moses tokenization script may not work well for Indian language
scripts. You can simply tokenize by space and puncutuations instead of
using the Moses Tokenizer.
A trivial tokenizer for Indian languages can be found here:
https://github.com/anoopkunchukuttan/indic_nlp_library
Regards
Anoop.
--------------------
Date: Tue, 18 Nov 2014 00:58:00 -0800
From: shiva kumar <shivadvg19@yahoo.com>
Subject: [Moses-support] normalization issue in tokenization of
Kannada words in baseline MT
To: moses-support@mit.edu
Message-ID:
<1416301080.63798.YahooMailBasic@web162302.mail.bf1.yahoo.com>
Content-Type: text/plain; charset=us-ascii
hi
i am working on baseline SMT with moses for Kannada-english MT. in the
tokenization step the input unicode fonts of kannada words will get added
with their unicode references because of glyph substitution.
due to this i am not able to get good translation. if i give the tokenized
sentences as input to decoder i am getting correct translation.
how to solve this problem?
i am using ubuntu12.04 and moses.
regards,
ShivaKumar KM
Asst.Professor,
Amrita VishwaVidyaPeetham Mysore Campus
Bogadi 2nd stage
Mysore
9611913393
--
I claim to be a simple individual liable to err like any other fellow
mortal. I own, however, that I have humility enough to confess my errors
and to retrace my steps.
http://flightsofthought.blogspot.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20141121/55dc4e1f/attachment-0001.htm
------------------------------
_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support
End of Moses-support Digest, Vol 97, Issue 66
*********************************************
Subscribe to:
Post Comments (Atom)
0 Response to "Moses-support Digest, Vol 97, Issue 66"
Post a Comment