Moses-support Digest, Vol 99, Issue 23

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: nplm Building LM (Nikolay Bogoychev)
2. Re: Feature score deltas in the chart decoder (Hieu Hoang)
3. CfP: 2nd Workshop on Multi-word Units in Machine Translation
and Translation Technology (MUMTTT 2015)
(MONTI JOHANNA -Professore associato scienze umanistiche e sociali-d)


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

Message: 1
Date: Mon, 12 Jan 2015 09:51:21 +0000
From: Nikolay Bogoychev <nheart@gmail.com>
Subject: Re: [Moses-support] nplm Building LM
To: Marwa Refaie <basmallah@hotmail.com>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAJzPUEy75jo700KVjw-GfvvddcJHnPMsVXb+XvnnZgdr3tEcKQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hey,

Refer to the moses documentation on how to use NPLM LM during decoding:
http://www.statmt.org/moses/?n=FactoredTraining.BuildingLanguageModel#ntoc31
In particular you need to add this:

NeuralLM factor=<factor> order=<order> path=filename


To your moses.ini where filename is model.NUMBER.

The 10 files model.1, model.2 etc are the neural network LM output after
each iteration/generation of training. So model.1 is the first generation
and model.10 is the 10th generation.

Cheers,

Nick


On Mon, Jan 12, 2015 at 12:05 AM, Marwa Refaie <basmallah@hotmail.com>
wrote:

>
>
>
>
> Hi
>
> Please I need any step by step tutorial for the nplm .
> I compiled the package & make trainnueralnetworkLM , then I got the
> validation.ngrams & train.ngrams then I got 10 files , model.1 model.2
> ........... model.10.
>
> I ran the ./bjam --with nplm ....
>
> Then what to do next now ??
>
> Please any help ??
>
> *Marwa N. Refaie*
>
>
> _______________________________________________
> 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/20150112/aabeeef4/attachment-0001.htm

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

Message: 2
Date: Mon, 12 Jan 2015 11:48:02 +0000
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Feature score deltas in the chart decoder
To: Jun-ya NORIMATSU <info@jnory.com>, moses-support@mit.edu
Message-ID: <54B3B472.4040903@gmail.com>
Content-Type: text/plain; charset="utf-8"

thanks!

On 12/01/15 04:49, Jun-ya NORIMATSU wrote:
> Hi,
>
> > In the Moses master branch I found one other feature function that
> > requires modifications:
> >
> > moses/LM/DALMWrapper.cpp
> >
> > This feature is currently not covered by a regression test, and I don't
> > have any setup with this feature myself. I would not be able to test any
> > modifications in that code and therefore would like to request that the
> > authors apply the necessary updates themselves.
>
> I've just finished to check and update the code.
> https://github.com/moses-smt/mosesdecoder/commit/39799188a0478eda822167202fb7d404b35fbaad
> https://github.com/moses-smt/mosesdecoder/commit/832b725c595d586f8802be12588fce9b495d36b8
>
> Please let me know if you find some problems.
>
> By the way, I'd like to add the regression test in near future.
> Would you mind my asking the place(directories or repositories...) to
> add the test code?
You can download the regression tests and the data needed from
https://github.com/moses-smt/moses-regression-tests
Add your test to the 'tests' directory. Add your LM to the 'LM' directory.

I would make a copy of an existing test, eg. phrase.basic-surface-only,
and name it
phrase.basic-surface-only.withDALM
and change it to use DALM instead.

You can test only your new test by running
./bjam --with-DALM=... --with-regtest=....
*phrase.basic-surface-only.withDALM.passed*

>
> Thanks,
> --
> Jun-ya NORIMATSU
>
>> ??: [Moses-support] Feature score deltas in the chart decoder
>> ??: 2015-01-08 04:17
>> ???: Matthias Huck <mhuck@inf.ed.ac.uk>
>> ??: Moses-support <moses-support@mit.edu>
>>
>> Hi,
>>
>> I've just pushed a commit to Moses that brings about a slight change
>> wrt. the way the chart decoder deals with feature scores.
>>
>> The chart decoder now stores deltas of individual feature scores instead
>> of constantly summing everything up. This behaviour is similar to what
>> we have been doing in the phrase-based decoder since a long time
>> already. The main purpose of this modification is to improve efficiency
>> with sparse features a bit.
>>
>> https://github.com/moses-smt/mosesdecoder/commit/465b47566424efb707bdc063d0bff52b0650eb0a
>>
>>
>>
>> The modification may however break existing feature function
>> implementations.
>>
>> As a rule of thumb, any feature function that calls
>>
>> ScoreComponentCollection::Assign()
>> in
>> EvaluateWhenApplied(const ChartHypothesis&, ...)
>>
>> is affected and needs to be adapted to the new behaviour.
>>
>> Basically, the ScoreComponentCollection variable passed to
>> EvaluateWhenApplied() now accumulates the delta score of the current
>> rule application only, whereas it was previously accumulating the
>> overall score of the partial hypothesis.
>> I.e., calling Assign() in EvaluateWhenApplied() now does not replace the
>> overall score any more, but has the same effect as calling PlusEquals().
>>
>> If you are the author of a feature function that implements
>> EvaluateWhenApplied(const ChartHypothesis&, ...) and calls Assign()
>> within that method, or if you are using such a feature function in your
>> experiments, please update your implementation. The feature function
>> should call PlusEquals() instead and add a score delta.
>>
>> I've already updated moses/LM/Ken.cpp and moses/LM/Implementation.cpp
>> and Rico has updated moses/LM/BilingualLM.cpp .
>> In the Moses master branch I found one other feature function that
>> requires modifications:
>>
>> moses/LM/DALMWrapper.cpp
>>
>> This feature is currently not covered by a regression test, and I don't
>> have any setup with this feature myself. I would not be able to test any
>> modifications in that code and therefore would like to request that the
>> authors apply the necessary updates themselves.
>>
>> Please let me know in case you notice any issues or if you need any
>> further information or advice regarding this modification.
>>
>> Cheers,
>> Matthias
> _______________________________________________
> 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/20150112/7e8d23ba/attachment-0001.htm

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

Message: 3
Date: Mon, 12 Jan 2015 18:18:12 +0100
From: "MONTI JOHANNA -Professore associato scienze umanistiche e
sociali-d" <jmonti@uniss.it>
Subject: [Moses-support] CfP: 2nd Workshop on Multi-word Units in
Machine Translation and Translation Technology (MUMTTT 2015)
To: <corpora@uib.no>, <dbworld@cs.wisc.edu>, <elsnet-list@elsnet.org>,
<flarenet_subscribers@ilc.cnr.it>, <IRList@lists.shef.ac.uk>,
<LINGUIST@listserv.linguistlist.org>, <moses-support@mit.edu>,
<mt-list@eamt.org>, corpora@clu.bccs.uib.no,
elsnet-list@cogsci.ed.ac.uk, info@aitla.it,
info@societadilinguisticaitaliana.net, ln@frmop11.bitnet,
openlogos-list@dfki.de, sito.sig@gmail.com,
multiword-expressions@lists.sourceforge.net,
PARSEME-all@chopin.ipipan.waw.pl, siglex@aclweb.org
Message-ID: <20150104161317.M96071@uniss.it>
Content-Type: text/plain; charset="utf-8"

The 2nd Workshop onMulti-word Units in Machine Translation and Translation Technology (MUMTTT2015)

to be held as part of theEUROPHRAS2015: "Computerised and Corpus-based Approaches to Phraseology:Monolingual and Multilingual Perspectives" (M?laga, 29 June ? 1 July 2015)

?

?endorsed by EUROPHRAS, European Society of Phraseology;the ICT COST action Parsing and multi-word expressions. Towards linguisticprecision and computational efficiency in natural language processing (PARSEME); the Special Interest Group on theLexicon of the Association for Computational Linguistics (SIGLEX); and SIGLEX?s MultiwordExpressions Section (SIGLEX-MWE).?

?

Brief description of theworkshop topic and content

Following the success of theMT SUMMIT 2013 Workshop on Multi-word Units in Machine Translation andTranslation Technology, we are announcing the 2015 edition to be held inconjunction with the Europhras conference on Computerised and Corpus-basedApproaches to Phraseology: Monolingual and Multilingual Perspectives (Malaga,Spain, 29 June ? 1 July 2015). The Workshop will provide a forum forresearchers working in the fields of Linguistics, Computational Linguistics,Translation Studies and Computational Phraseology to discuss recent advances inthe area of multi-word unit processing and to coordinate research effortsacross disciplines.

In spite of the recentprogress achieved in Machine Translation and Translation Technology, theidentification, interpretation and translation of multi-word units (MWUs) stillrepresent open challenges, both from a theoretical and a practical point ofview. The idiosyncratic morpho-syntactic, semantic and translational propertiesof the linguistic phenomenon associated with the above tasks poses manyobstacles even to human translators, mainly because of intrinsic ambiguities,structural and lexical asymmetries between languages, and, finally, culturaldifferences.

In recent years, growingattention has been paid to integrating MWU in machine translation andtranslation technology tools, as it has been acknowledged that it is notpossible to create large scale applications without properly handling MWUs ofall kinds. As a matter of fact, researchers are now addressing the problemsposed by MWU processing and translation using different formalisms andtechniques, such as: automatic recognition of MWUs in a monolingual orbilingual setting; alignment and paraphrasing methodologies; development,features and usefulness of handcrafted monolingual and bilingual linguisticresources and grammars.

The aim of the Workshop is tobring together researchers working on various aspects of MWU processing indifferent disciplines, in order to discuss and propose innovative ideas andmethods in relation to Machine Translation and Translation Technology. Inparticular, this workshop welcomes the interactions between researchers inNatural Language Processing working on the computational treatment ofmulti-word units, experts in phraseology (including computational phraseology)working on challenging topics of their discipline, as well as translationpractitioners, to the benefit of applying their latest results to advance thestate of the art in MWU translation.

MUMTTT 2015 is the secondevent of a series dedicated to Multiword Unit processing and translation in thecontext of Machine Translation and Translation Technology. We are interested inoriginal research related (but not limited) to the following topics:

?????????Lexical, syntactic, semantic and translational aspectsin MWU representation

?????????Collostructional analysis and MWU

?????????Automatic extraction of multilingual MWU resources

?????????Identification and acquisition of multi-word terms andtheir variants

?????????Development and use of handcrafted MWU linguisticresources in MT

?????????Development of corpora for extraction and translationof MWUs

?????????MWUs in Computer-assisted Translation

?????????MWUs in Machine Translation

?????????Paraphrasing of MWUs applied to the improving of MT

?????????MWUs and word alignment techniques

?????????Learning semantic information about MWUs frommonolingual, parallel or comparable corpora

?????????Multilingualism and MWU processing

?????????MWU-centred MT evaluation

?????????Compilation of resources for extraction andtranslation of multiword units

?????????Creation of MWU-annotated corpora with a focus ontranslation aspects

?

?

Workshop Organisers

Gloria Corpas Pastor (Universidad de M?laga, Spain)

Ruslan Mitkov (University of Wolverhampton, UK)?

Johanna Monti (University of Sassar, Italy)

Violeta Seretan (University of Geneva, Switzerland)

?

Advisory Board

Dmitrij O. Dobrovol'skij (Russian Academy ofSciences, Russia)

Carlos Ramisch (Aix-Marseille University, France)

Michael Rosner (University of Malta)

Agata Savary (Universit? Fran?ois Rabelais Tours, France)

Kathrin Steyer, (Institut f?r Deutsche Sprache,Germany)

?

Abstract submission andimportant dates
* ? Abstract length: between400 and 500?words?(without bibliography).
* ?Deadline forabstract submission:?15th March 2015.
* ? Abstracts are to besubmitted through the conference online management system.
* ?Notification ofacceptance:?30th April 2015.
* ?Submission of fullpapers (according to style sheet):?15th July 2015.

?

Programme Committee members

I?aki Alegria (University of the BasqueCountry, Spain)

Giuseppe Attardi (University of Pisa, Italy)

Doug Arnold (University of Essex, UnitedKingdom)

Ant?nio Branco (University of Lisbon, Portugal)

Paul Buitelaar (National University of Ireland,Galway)

Jean-Pierre Colson (Universit? catholique de Louvain, Belgium)

Franti?ek ?erm?k (Charles University in Prague,Czech Republic)

Matthieu Constant (Universit? Paris-Est, France)

Ga?l Dias (University of Caen Basse-Normandie,France)

Mike Dillinger ( Association for MachineTranslation in the Americas)

Dmitrij O. Dobrovol'skij (Russian Academy ofSciences, Russia)

Peter ?ur?o (University of Cyril and Method inTrnava, Slovak Republic)

Marcello Federico (FBK, Italy)

Sabine Fiedler (University of Leipzig, Germany)

Natalia Filatkina (University of Trier,Germany)

Thierry Fontenelle (Translation Centre for theBodies of the European Union, Luxembourg)

Corina For?scu (Al.I. Cuza University of Iasi, Romania)

Thomas Fran?ois (Universite catholique de Louvain, Belgium)

Ulrich Heid (Universit?t Hildesheim, Germany)

Kyo Kageura (University of Tokyo, Japan)

Cvetana Krstev (University of Belgrade, Serbia)

Koenraad Kuiper (University of Canterbury, NewZealand)

Alon Lavie (Carnegie Mellon University, USA)

Malvina Nissim (University of Bologna, Italy)

Michael Oakes (University of Wolverhampton,United Kingdom)

Preslav Nakov (Qatar Computing ResearchInstitute, Qatar Foundation, Qatar)

Adriana Orlandi (University of Modena and Reggio Emilia,Italy)

Yannick Parmentier (University of Orl?ansFrance)

Pavel Pecina (Charles University in Prague,Czech Republic)

Carlos Ramisch (University of Grenoble, France)

Victoria Ros?n (University of Bergen, Norway)

Michael Rosner (University of Malta)

Manfred Sailer (Goethe Universit?t, Germany)

Agata Savary (Universit? Fran?ois Rabelais Tours, France)

Tanja Samard?i? (University of Zurich, Switzerland)

Gerold Schneider (University of Zurich, Switzerland)

Gilles S?rasset (University of Grenoble, France)

Max Silberztein (University of Franche-Comt?, France)

Kiril Simov (Bulgarian Academy of Sciences,Bulgaria)

Kathrin Steyer, (Institut f?r Deutsche Sprache,Germany)

Joanna Szerszunowicz (University of Bialystok,Poland)

Marko Tadi? (University of Zagreb, Crotia)

Amalia Todirascu (Universit? de Strasbourg, France)

Beata Trawinski (Institut f?r Deutsche SpracheMannheim, Germany)

Dan Tufi? (Romanian Academy, Romania)

Agn?s Tutin (University of Grenoble, France)

Michael Zock (Aix-Marseille University, France)

Lonneke van der Plas (University of Stuttgart,Germany)

Veronika Vincze (University of Szeged, Hungary)

Martin Volk (University of Zurich, Switzerland)

Eric Wehrli (Universit? de Gen?ve, Switzerland)

Contact

Forany inquiries regarding the workshop please send an email to: jmonti@uniss.it orgcorpas@uma.es.

PARSEME grants

PARSEME will fund the travel and stay at the MUMTTT 2015 workshop forabout 20 participants from the PARSEME member countries. For any information please refer to:? http://typo.uni-konstanz.de/parseme/index.php/events/118-mumttt-workshop-at-europhras-15

Call

Send SMS

Add to Skype

You'll need Skype CreditFree via Skype


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150112/85d2fbc1/attachment.htm

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

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


End of Moses-support Digest, Vol 99, Issue 23
*********************************************

Tweet