Moses-support Digest, Vol 104, Issue 71

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. Re: Exception when exiting moses (Hieu Hoang)
2. BLEU Score Variance: Which score to use? (Hokage Sama)
3. Invitation-based data selection tool for bilingual domain
adaptation (Amir Kamran)
4. Re: BLEU Score Variance: Which score to use?
(Marcin Junczys-Dowmunt)


----------------------------------------------------------------------

Message: 1
Date: Sun, 21 Jun 2015 21:29:27 +0400
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Exception when exiting moses
To: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAEKMkbjKNits0rStJhCazZZb9StOyLR+kbn65mR2PT2GBQ1q5Q@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

cheers big ears

Hieu Hoang
Researcher
New York University, Abu Dhabi
http://www.hoang.co.uk/hieu

On 21 June 2015 at 21:25, Marcin Junczys-Dowmunt <junczys@amu.edu.pl> wrote:

> Forgot to remove C++11 specific stuff. Try now.
>
> On 21.06.2015 19:17, Hieu Hoang wrote:
>
>> i'm getting a compile error on my mac with clang
>>
>>
>> Hieu Hoang
>> Researcher
>> New York University, Abu Dhabi
>> http://www.hoang.co.uk/hieu
>>
>> On 21 June 2015 at 21:07, Jani Dugonik <Jani.Dugonik@um.si <mailto:
>> Jani.Dugonik@um.si>> wrote:
>>
>> Yes, its working now.
>>
>> Thanks,
>> Jani
>> ________________________________________
>> Od: moses-support-bounces@mit.edu
>> <mailto:moses-support-bounces@mit.edu>
>> [moses-support-bounces@mit.edu
>> <mailto:moses-support-bounces@mit.edu>] v imenu Marcin
>> Junczys-Dowmunt [junczys@amu.edu.pl <mailto:junczys@amu.edu.pl>]
>> Poslano: 21. junij 2015 18:28
>> Za: moses-support@mit.edu <mailto:moses-support@mit.edu>
>> Zadeva: Re: [Moses-support] Exception when exiting moses
>>
>> You can check now.
>>
>> On 21.06.2015 18:25, Marcin Junczys-Dowmunt wrote:
>> > I think I have it now, running a few tests and then push. Not
>> sure what
>> > actually caused it in combination with C++11, but some general code
>> > cleaning helped.
>> >
>> > On 21.06.2015 18:14, Jani Dugonik wrote:
>> >> Yes, I'm also having the problem when running MERT, here are
>> the last lines:
>> >>
>> >> terminate called after throwing an instance of
>> 'util::ErrnoException'
>> >> what(): util/mmap.cc:52 in void util::UnmapOrThrow(void*,
>> size_t) threw ErrnoException because `munmap(start, length)'.
>> >> Invalid argument munmap failed
>> >> Aborted
>> >> Exit code: 134
>> >> The decoder died.
>> >>
>> >> I hope this helps,
>> >> Jani
>> >>
>> > _______________________________________________
>> > Moses-support mailing list
>> > Moses-support@mit.edu <mailto:Moses-support@mit.edu>
>> > http://mailman.mit.edu/mailman/listinfo/moses-support
>>
>> _______________________________________________
>> Moses-support mailing list
>> Moses-support@mit.edu <mailto:Moses-support@mit.edu>
>> http://mailman.mit.edu/mailman/listinfo/moses-support
>>
>> _______________________________________________
>> Moses-support mailing list
>> Moses-support@mit.edu <mailto: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/20150621/6ddfbafe/attachment-0001.htm

------------------------------

Message: 2
Date: Sun, 21 Jun 2015 19:37:16 -0500
From: Hokage Sama <nvncbol@gmail.com>
Subject: [Moses-support] BLEU Score Variance: Which score to use?
To: moses-support <moses-support@mit.edu>
Message-ID:
<CAD3ogMbpj3pef6FVrB1Qfe-FuAQ91dvqXA7SJ3eQgY4VPp65Ng@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hi,

Since MT training is non-convex and thus the BLEU score varies, which score
should I use for my system? I trained my system three times using the same
data and obtained the three different scores below. Should I take the
average or the best score?

BLEU = 17.84, 49.1/22.0/12.5/7.5 (BP=1.000, ratio=1.095, hyp_len=3952,
ref_len=3609)
BLEU = 16.51, 48.4/20.7/11.4/6.5 (BP=1.000, ratio=1.093, hyp_len=3945,
ref_len=3609)
BLEU = 15.33, 48.2/20.1/10.3/5.5 (BP=1.000, ratio=1.087, hyp_len=3924,
ref_len=3609)

Thanks,
Hilton
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150621/f0ab5b35/attachment-0001.htm

------------------------------

Message: 3
Date: Mon, 22 Jun 2015 07:45:15 +0200
From: Amir Kamran <amirkamran@gmail.com>
Subject: [Moses-support] Invitation-based data selection tool for
bilingual domain adaptation
To: moses-support@mit.edu
Message-ID:
<CAL0aJKiXA6pXBSAyEx+sCDQXszTOc0-E4eup5c6n3hJss12KeA@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Dear All,

I implemented a data selection tool for domain adaptation based on
Invitation Model as described in:
Hoang, Cuong and Sima'an, Khalil (2014): Latent Domain Translation Models
in Mix-of-Domains Haystack, Proceedings of COLING 2014, the 25th
International Conference on Computational Linguistics,
http://www.aclweb.org/anthology/C14-1182.pdf

The developed tool is available at the following Github repository:

*https://github.com/amirkamran/InvitationModel*
<https://github.com/amirkamran/InvitationModel>


Invitation based data selection approach exploits in-domain data (both
monolingual and bilingual) as prior to guide word alignment and phrase pair
estimates in the large mix-domain corpus. As a by-product, accurate
estimates for P(D|e,f) of the mixed-domain sentences are produced (with D
being either in-domain or out-of-domain), which can be used to rank the
sentences in mix-domain according to their relevance to in-domain corpus.

This work has been conducted at ILLC (Institute for Logic, Language and
Computation, University of Amsterdam) https://www.illc.uva.nl as part of
the project "Data-Powered Domain-Specific Translation Services On Demand",
supported by the grant "STW Open Technologieprogramma".


Regards
Amir Kamran
Research Programmer
Institute of Logic, Language and Computation (ILLC)
University of Amsterdam
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150622/6362f998/attachment-0001.htm

------------------------------

Message: 4
Date: Mon, 22 Jun 2015 08:00:25 +0200
From: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Subject: Re: [Moses-support] BLEU Score Variance: Which score to use?
To: moses-support@mit.edu
Message-ID: <5587A479.7090201@amu.edu.pl>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

Hi,
I think the average is OK, your variance is however quite high. Did you
retrain the entire system or just optimize parameters a couple of times?

Two useful papers on the topic:

https://www.cs.cmu.edu/~jhclark/pubs/significance.pdf
http://www.mt-archive.info/MTS-2011-Cettolo.pdf


On 22.06.2015 02:37, Hokage Sama wrote:
> Hi,
>
> Since MT training is non-convex and thus the BLEU score varies, which
> score should I use for my system? I trained my system three times
> using the same data and obtained the three different scores below.
> Should I take the average or the best score?
>
> BLEU = 17.84, 49.1/22.0/12.5/7.5 (BP=1.000, ratio=1.095, hyp_len=3952,
> ref_len=3609)
> BLEU = 16.51, 48.4/20.7/11.4/6.5 (BP=1.000, ratio=1.093, hyp_len=3945,
> ref_len=3609)
> BLEU = 15.33, 48.2/20.1/10.3/5.5 (BP=1.000, ratio=1.087, hyp_len=3924,
> ref_len=3609)
>
> Thanks,
> Hilton
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support



------------------------------

_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support


End of Moses-support Digest, Vol 104, Issue 71
**********************************************

Tweet