Moses-support Digest, Vol 92, Issue 33

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: Feature function with variable number of scores (Hieu Hoang)
2. AMTA 2014 Research Papers: NEW EXTENDED SUBMISSION DEADLINE
(Simard, Michel)
3. (no subject) (Kalyani Baruah)


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

Message: 1
Date: Sat, 14 Jun 2014 20:36:27 +0100
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] Feature function with variable number of
scores
To: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Cc: Hieu Hoang <hieu.hoang@ed.ac.uk>, moses-support
<moses-support@mit.edu>
Message-ID: <ED56DB7D-F732-4A83-AFF0-FCF5C79020E8@gmail.com>
Content-Type: text/plain; charset="us-ascii"

That'll work

We could add a function to reset the number of scores but it's not a big priority

Sent from my flying horse

> On 14 Jun 2014, at 20:22, Marcin Junczys-Dowmunt <junczys@amu.edu.pl> wrote:
>
> Came up with something like that, still ugly, but works. num-features can be specified too, and it will complain if it does not match the length of "scores"
>
> std::string ParseScores(const std::string &line, const std::string& defaultScores) {
> vector<string> toks = Tokenize(line);
> UTIL_THROW_IF2(toks.empty(), "Empty line");
>
> for (size_t i = 1; i < toks.size(); ++i) {
> std::vector<std::string> args = TokenizeFirstOnly(toks[i], "=");
> UTIL_THROW_IF2(args.size() != 2,
> "Incorrect format for feature function arg: " << toks[i]);
>
> if (args[0] == "scores") {
> return args[1];
> }
> }
> return defaultScores;
> }
>
> EditOps::EditOps(const std::string &line)
> : StatelessFeatureFunction(ParseScores(line, "mdis").size(), line),
> m_factorType(0), m_chars(false), m_scores(ParseScores(line, "mdis"))
> {
> std::cerr << "Initializing EditOps feature.." << std::endl;
> ReadParameters();
> }
>
>
> W dniu 14.06.2014 20:28, Hieu Hoang pisze:
>>
>>
>>
>>> On 14 June 2014 19:19, Marcin Junczys-Dowmunt <junczys@amu.edu.pl> wrote:
>>> Actually, I do know it at construction time. For instance I specify a string "scores=mdis" in the ini, every character in the value corresponds to one type of score, so I have 4 feature scores. Setting again "num-features=4" is redundant now (I guess I can still do that, but I don't like it). I see the number of scores is set in the low-level function ParseLine in the FeatureFunction class, should I override ParseLine in the inherited class then?
>> let's not override ParseLine yet. I can't see what's wrong with
>> num-features=4
>> We need some kind of consistency
>>
>>> W dniu 14.06.2014 20:11, Hieu Hoang pisze:
>>>
>>>> the number of scores must be set in the feature function constructor, there's no functionality to change it once its set
>>>>
>>>> You might be better off using sparse features if you don't know how many scores you have at construction time
>>>>
>>>>
>>>>> On 14/06/14 09:17, Marcin Junczys-Dowmunt wrote:
>>>>> Hi,
>>>>> what's the trick for defining a feature function with a variable number
>>>>> of scores based on custom parameters in the ini-file?
>>>>> Just setting m_numScoreComponents = 2 in SetParameter(...) doesn't work.
>>>>> It seems the number of features is registered somewhere on object
>>>>> creation as I get the followiing message:
>>>>>
>>>>> Exception: moses/ScoreComponentCollection.cpp:242 in void
>>>>> Moses::ScoreComponentCollection::Assign(const Moses::FeatureFunction*,
>>>>> const std::vector<float>&) threw util::Exception'.
>>>>> Feature function EditOps0 specified 1 dense scores or weights. Actually
>>>>> has 2
>>>>>
>>>>> Thanks,
>>>>> Marcin
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Moses-support mailing list
>>>>> Moses-support@mit.edu
>>>>> http://mailman.mit.edu/mailman/listinfo/moses-support
>>
>>
>> --
>> Hieu Hoang
>> Research Associate
>> University of Edinburgh
>> http://www.hoang.co.uk/hieu
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140614/04fd573d/attachment-0001.htm

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

Message: 2
Date: Fri, 13 Jun 2014 08:47:21 -0400
From: "Simard, Michel" <Michel.Simard@cnrc-nrc.gc.ca>
Subject: [Moses-support] AMTA 2014 Research Papers: NEW EXTENDED
SUBMISSION DEADLINE
To: "corpora@uib.no" <corpora@uib.no>, "mt-list@eamt.org"
<mt-list@eamt.org>, "moses-support@mit.edu" <moses-support@MIT.EDU>,
"sigmt@aclweb.org" <sigmt@aclweb.org>
Cc: "Yaser Al-onaizan \(onaizan@us.ibm.com\)" <onaizan@us.ibm.com>,
"Foster, George" <george.foster@cnrc-nrc.gc.ca>,
"mike.dillinger@translationoptimization.com"
<mike.dillinger@translationoptimization.com>
Message-ID:
<D7548FA9B5763F408F5EB57EE28383621B1ED74908@NRCCENMB1.nrc.ca>
Content-Type: text/plain; charset="us-ascii"

AMTA 2014 FINAL CALL FOR PAPERS

**** NEW EXTENDED SUBMISSION DEADLINE: MONDAY JUNE 23, 2014 ****

AMTA-2014 solicits original research papers that will advance the field of Machine Translation. We seek submissions across the entire spectrum of MT-related research activity. Submissions must be unpublished, and in English.

IMPORTANT DATES:

Submission deadline: Monday, June 23
Notification of acceptance: Monday, August 4
Final "camera-ready" versions: Monday, September 8

TOPICS OF INTEREST

Topics of interest include but are not limited to:

- Advances in various MT paradigms: data-driven, rule-based, and hybrids

- MT applications and embedding: translation/localization aids, speech-to-speech, speech-to-text, OCR, MT for communication (chats, blogs, social networks), multilingual applications, etc.

- Technologies for MT deployment: quality estimation and domain adaptation

- MT in special settings: low resources, massive resources, high volume, low computing resources

- Human factors in MT and user interfaces for MT

- Linguistic resources for MT: dictionaries, terminology banks, corpora

- MT evaluation techniques and evaluation results

- Empirical studies on translation data

SUBMISSION INSTRUCTIONS:

Papers must not exceed 12 (twelve) pages plus 4 (four) pages for references. All papers should follow the formatting instructions included with the style files, and should be submitted in PDF. Latex, PDF and MS Word style files are available: AMTA2014.latex.zip amta2014.dotx amta2014.pdf

To allow for blind reviewing, please do not include author names and affiliations within the paper and avoid obvious self-references.

Papers must be submitted to the START system by 11:59 pm PDT (GMT - 7 hours), Monday, June 23, 2014.

MULTIPLE SUBMISSIONS

Papers presented at the AMTA conference must represent new work that has not been previously published. It is the responsiblity of the author to inform the program chairs of any potential problem with respect to this requirement. Authors submitting a similar paper both to AMTA and another conference or workshop must inform the program chairs by email, specifying to which other conference or workshop they are submitting their work. If the paper is presented at the AMTA conference, then it must be withdrawn from other conferences and workshops.


--
Michel Simard
Research Officer
National Research Council Canada




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

Message: 3
Date: Sat, 14 Jun 2014 18:14:59 -0700
From: Kalyani Baruah <kajubaruah04@gmail.com>
Subject: [Moses-support] (no subject)
To: "moses-support@mit.edu" <moses-support@mit.edu>,
moses-support@gmail.com
Message-ID:
<CAJZ5LDdKXFDRVwNQPneA2-AfpiAgTuGPAajLen+u-av-LZfBAQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hiee Good day everyone.. English words have only one connection because
they dont have any inflection is 'int it.am i correct.and using BLEU and
WER,PER gives a vast difference why is it so.

Regards,



*Kalyanee Kanchan Baruah*
Institute of Science and Technology,
Gauhati University,Guwahati,India
Phone- +91-9706242124
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20140614/23ff337a/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 92, Issue 33
*********************************************

Tweet