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. --activate-features in mert-moses.perl not working?
(Marcin Junczys-Dowmunt)
2. Deadline extension: 9th SaLTMiL workshop on ?Free/open-source
language resources for the machine translation of less-resourced
languages? at LREC 2014 (Mikel Forcada)
3. Re: --activate-features in mert-moses.perl not working?
(Barry Haddow)
4. Re: --activate-features in mert-moses.perl not working?
(Marcin Junczys-Dowmunt)
----------------------------------------------------------------------
Message: 1
Date: Mon, 10 Feb 2014 19:01:14 +0100
From: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Subject: [Moses-support] --activate-features in mert-moses.perl not
working?
To: moses-support@MIT.EDU
Message-ID: <52F913EA.9040009@amu.edu.pl>
Content-Type: text/plain; charset=UTF-8; format=flowed
Hi,
it seems --activate-features=STRING is not working in mert-moses.perl.
The script prints a message that the ignored features are not being
used, but then optimizes them anyway. I can see that the "enabled"
information in the feature data structure is not being used anywhere in
the script once it has been set (apart from printing the message).
This can cause an interesting catastrophe when, for instance, distortion
is disabled by setting the limit to 1:
MERT assigns a weight of 1 to distortion (but the feature itself is
always 0) and 0 weights to all other features, the final score is then
equal to 0 for all sentences and poor moses goes crazy generating lots
of garbage which in turn takes ages to score only to finish with bad
weights. Really ugly, took me a while to find the cause :)
BTW. In my opinon a --deactive-features might be more useful. I would
add/correct it myself, but currently I am getting lost in the code that
is printing the config files. Someone more acquainted with that code?
Best,
Marcin
------------------------------
Message: 2
Date: Mon, 10 Feb 2014 20:01:55 +0100
From: Mikel Forcada <mlf@dlsi.ua.es>
Subject: [Moses-support] Deadline extension: 9th SaLTMiL workshop on
?Free/open-source language resources for the machine translation of
less-resourced languages? at LREC 2014
To: moses-support@mit.edu
Message-ID: <52F92223.6040901@dlsi.ua.es>
Content-Type: text/plain; charset="windows-1252"
***** NEW DEADLINE: February 17 *****
Call for Papers: 9th SaLTMiL workshop on ?Free/open-source language
resources for the machine translation of less-resourced languages? at
LREC 2014
A full-day workshop at LREC 2014
Tuesday, 27 May 2014.
Reykjavik (Iceland)
SALTMIL: http://ixa2.si.ehu.es/saltmil/
LREC 2014: http://lrec2014.lrec-conf.org/en/
Website: http://ixa2.si.ehu.es/saltmil/
Paper submission: https://www.softconf.com/lrec2014/SaLTMiL/
The 9th International Workshop of the Special Interest Group on Speech
and Language Technology for Minority Languages (SaLTMiL) will be held in
Reykjav?k, Iceland, on May 24, 2014, as part of the 2014 International
Language Resources and Evaluation Conference (LREC). (For SALTMIL see:
http://ixa2.si.ehu.es/saltmil/); it is also framed as one of the
activities of European project Abu-Matran (http://www.abumatran.eu).
Entitled "Free/open-source language resources for the machine
translation of less-resourced languages", the workshop is intended to
continue the series of SALTMIL/LREC workshops on computational language
resources for minority languages, held in Granada (1998), Athens (2000),
Las Palmas de Gran Canaria (2002), Lisbon (2004), Genoa (2006),
Marrakech (2008), La Valetta (2010) and Istanbul (2012), and is also
expected to attract the audience of Free Rule-Based Machine Translation
workshops (2009, 2011, 2012). The workshop aims to share information on
language resources, tools and best practice, to save isolated
researchers from starting from scratch when building machine translation
for a less-resourced language. An important aspect will be the
strengthening of the free/open-source language resources community,
which can minimize duplication of effort and optimize development and
adoption, in line with the LREC 2014 hot topic ?LRs in the Collaborative
Age? (http://is.gd/LREChot).
The whole-day workshop will consist of short oral papers, a poster
session preceded by a poster-boaster session (2 minutes, 2 slides per
poster), and a round table.
Papers are invited that describe research and development in the
following areas:
FOS LR for rule-based machine translation (dictionaries, rule sets)
FOS LR for statistical machine translation (corpora)
FOS tools to annotate, clean, preprocess, convert, etc. LRs for machine
translation
Machine translation as a tool for creating or enriching FOS LRs for
less-resourced languages
Position papers and (web based) demonstrations will also be considered
for presentation.
The best papers, as evaluated by the programme committee, will be
presented orally and the remaining paper will be presented in poster
format.
We expect short papers of max 6,000 words (up to 6 pages) describing
research addressing one of the above topics, to be submitted as PDF
documents by using the LREC 2014 START conference management system
(https://www.softconf.com/lrec2014/SaLTMiL/).
Submissions should be anonymized. When submitting a paper through the
START page, authors will be kindly asked to share the resources that
have been used for the work described in their paper or that are the
outcome of their research. For further information on this initiative,
please refer to
http://lrec2014.lrec-conf.org/en/calls-for-papers/lrec-2014-special-highlight/.
Submissions of papers should follow the same style as the papers for the
main LREC conference (an Author's Kit made of specific guidelines and
downloadable templates will be published on the conference web site in
due time). All contributions will be included in the workshop
proceedings (CD). They will also be published on the SALTMIL website.
The registration fees will be duly announced at the LREC 2014 site.
Registration in the workshop willl include a coffee break and the
Proceedings of the Workshop. Registration will be handled by the LREC
2014 Secretariat.
Important dates
Deadline for paper submission: February 10, 2014 *February 17, 2014*
Notification of acceptance sent: March, 3, 2014 *March 10, 2014*
Camera-ready paper due: March 21, 2014
Organizing committee
Joint e-mail address: saltmil2014@dlsi.ua.es
(1) Dr Francis M Tyers
Institutt for spr?kvitskap
Det humanistiske fakultet,
N-9037 Universitetet i Troms?
ftyers@prompsit.com
(2) Dr Kepa Sarasola
Computer Science Faculty
Dept. of Computer Languages
The University of the Basque Country
P.K. 649 20080 DONOSTIA
Basque Country, Spain
Tel: +34 943 01 81 54
Fax: +34 943 21 93 06
ksarasola@ehu.es
http://ixa.si.ehu.es
(3) Prof Mikel L. Forcada
Dept. Llenguatges i Sistemes inform?tics
Universitat d?Alacant
E-03071 Alacant (Spain)
Tel: +34 96 590 9776
FAx: +34 96 590 9326
mlf@ua.es
http://www.dlsi.ua.es/~mlf
Programme Committee
I?aki Alegria, Euskal Herriko Unibertsitatea, Spain
Lars Borin, G?teborgs Universitet, Sweden.
Elaine U? Dhonnchadha, Trinity College Dublin, Ireland
Mikel L. Forcada, Universitat d?Alacant, Spain
Michael Gasser, Indiana University, USA
M?ns Huld?n, Helsingin Yliopisto, Finland
Krister Lind?n, Helsingin Yliopisto, Finland
Nikola Ljube?ic', Sveuc(ili?te u Zagrebu, Croatia
Llu?s Padr?, Universitat Polit?cnica de Catalunya, Spain
Juan Antonio P?rez-Ortiz, Universitat d?Alacant, Spain
Felipe S?nchez-Mart?nez, Universitat d?Alacant
Kepa Sarasola, Euskal Herriko Unibertsitatea, Spain
Kevin P. Scannell, Saint Louis University, USA
Antonio Toral, Dublin City University, Ireland
Trond Trosterud, Universitet i Troms?, Norway
Francis M. Tyers, Universitet i Troms?, Norway
--
Mikel L. Forcada (http://www.dlsi.ua.es/~mlf/)
Departament de Llenguatges i Sistemes Inform?tics
Universitat d'Alacant
E-03071 Alacant, Spain
Phone: +34 96 590 9776
Fax: +34 96 590 9326
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140210/2f3c9ac3/attachment-0001.htm
------------------------------
Message: 3
Date: Mon, 10 Feb 2014 19:46:17 +0000
From: Barry Haddow <bhaddow@staffmail.ed.ac.uk>
Subject: Re: [Moses-support] --activate-features in mert-moses.perl
not working?
To: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>, moses-support@mit.edu
Message-ID: <52F92C89.4000203@staffmail.ed.ac.uk>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Hi Marcin
I had some fun with --activate-features in the past - I think the syntax
was rather strange. If it is not working now, it may have got dropped by
the recent refactoring
My advice would be to use kbmira (or pro), since they are regularised
they don't go crazy when there is an uninformative feature. That way,
you don't have to fiddle with feature activation,
cheers - Barry
On 10/02/14 18:01, Marcin Junczys-Dowmunt wrote:
> Hi,
> it seems --activate-features=STRING is not working in mert-moses.perl.
> The script prints a message that the ignored features are not being
> used, but then optimizes them anyway. I can see that the "enabled"
> information in the feature data structure is not being used anywhere in
> the script once it has been set (apart from printing the message).
>
> This can cause an interesting catastrophe when, for instance, distortion
> is disabled by setting the limit to 1:
> MERT assigns a weight of 1 to distortion (but the feature itself is
> always 0) and 0 weights to all other features, the final score is then
> equal to 0 for all sentences and poor moses goes crazy generating lots
> of garbage which in turn takes ages to score only to finish with bad
> weights. Really ugly, took me a while to find the cause :)
>
> BTW. In my opinon a --deactive-features might be more useful. I would
> add/correct it myself, but currently I am getting lost in the code that
> is printing the config files. Someone more acquainted with that code?
> Best,
> Marcin
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
------------------------------
Message: 4
Date: Mon, 10 Feb 2014 20:59:01 +0100
From: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Subject: Re: [Moses-support] --activate-features in mert-moses.perl
not working?
To: moses-support@mit.edu
Message-ID: <52F92F85.8080301@amu.edu.pl>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
W dniu 10.02.2014 20:46, Barry Haddow pisze:
Ah, by the way, is removing the Distortion feature from the ini file and
setting the limit to 1 a safe way to actually disable distortion? Moses
does not complain (I always thought it is required.)
Best,
Marcin
> Hi Marcin
>
> I had some fun with --activate-features in the past - I think the
> syntax was rather strange. If it is not working now, it may have got
> dropped by the recent refactoring
>
> My advice would be to use kbmira (or pro), since they are regularised
> they don't go crazy when there is an uninformative feature. That way,
> you don't have to fiddle with feature activation,
>
> cheers - Barry
>
> On 10/02/14 18:01, Marcin Junczys-Dowmunt wrote:
>> Hi,
>> it seems --activate-features=STRING is not working in mert-moses.perl.
>> The script prints a message that the ignored features are not being
>> used, but then optimizes them anyway. I can see that the "enabled"
>> information in the feature data structure is not being used anywhere in
>> the script once it has been set (apart from printing the message).
>>
>> This can cause an interesting catastrophe when, for instance, distortion
>> is disabled by setting the limit to 1:
>> MERT assigns a weight of 1 to distortion (but the feature itself is
>> always 0) and 0 weights to all other features, the final score is then
>> equal to 0 for all sentences and poor moses goes crazy generating lots
>> of garbage which in turn takes ages to score only to finish with bad
>> weights. Really ugly, took me a while to find the cause :)
>>
>> BTW. In my opinon a --deactive-features might be more useful. I would
>> add/correct it myself, but currently I am getting lost in the code that
>> is printing the config files. Someone more acquainted with that code?
>> Best,
>> Marcin
>> _______________________________________________
>> 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 88, Issue 19
*********************************************
Subscribe to:
Post Comments (Atom)
0 Response to "Moses-support Digest, Vol 88, Issue 19"
Post a Comment