Moses-support Digest, Vol 85, Issue 35

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: Moses-support post from colin.a.cherry@gmail.com requires
approval (Colin Cherry)
2. Re: Moses-support post from colin.a.cherry@gmail.com requires
approval (Hieu Hoang)
3. CALL FOR PAPERS: Workshop on Interactive Language Learning,
Visualization, and Interfaces (Philipp Koehn)


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

Message: 1
Date: Wed, 20 Nov 2013 13:47:33 -0500
From: Colin Cherry <colin.a.cherry@gmail.com>
Subject: Re: [Moses-support] Moses-support post from
colin.a.cherry@gmail.com requires approval
To: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAFfbXc7K_5W=mo-MvZ4GU42dAM3b3oMJeiPrZUEMtUnTCMmmLw@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Thanks for getting back so quickly.

For the most part, things are working. For example, I can decode existing
models with no problems, and the training script that produces phrase
tables also appears to work. Unfortunately, for parameter tuning, extractor
appears to not be working. In particular, it is coming back with scores.dat
files that look like this:

SCORES_TXT_BEGIN_0 0 100 9 BLEU
0 0 0 0 0 0 0 0 2
0 0 0 0 0 0 0 0 2
0 0 0 0 0 0 0 0 2
0 0 0 0 0 0 0 0 2
0 0 0 0 0 0 0 0 2

This seems to be directly related to the MertBleuScorer test suite, which
is also failing by returning too many 0's, but that might be a red herring.
If no one else is observing this issue, then that's useful information. I
may look into MertBleuScorer's problems anyway, if I get a moment.

-- Colin


On Wed, Nov 20, 2013 at 11:47 AM, Hieu Hoang <Hieu.Hoang@ed.ac.uk> wrote:

> hi colin,
>
> I think i've fixed everything that needed to be fixed for clang
> compatibility. The unit tests never worked for me on Mac OSX anyway,
> something to do with boost static libraries.
>
> Your build.log looks like everything compiled, but the unit tests failed,
> like mine.
>
> The few times I've used my mac to train and tune, this was ok.
>
>
>> ---------- Forwarded message ----------
>> From: Colin Cherry <colin.a.cherry@gmail.com>
>> To: Moses-support@mit.edu
>> Cc:
>> Date: Wed, 20 Nov 2013 09:55:59 -0500
>> Subject: Trouble building on OS X 10.9 (Mavericks)
>> Hi all,
>>
>> I know Hieu has been working on this, but I'm still having some trouble
>> getting Moses to build on my Mac after upgrading to Mavericks. In
>> particular, extractor does not seem to be working properly, and relevant
>> unit tests in MertBleuScorer are failing.
>>
>> I've attached my build.log as suggested by the build process. Please let
>> me know if any other information would be helpful.
>>
>> -- Colin
>>
>>
>> ---------- Forwarded message ----------
>> From: moses-support-request@mit.edu
>> To:
>> Cc:
>> Date:
>> Subject: confirm 930e0cc9b7909a1ac82c7494444520fd44d8bf94
>> If you reply to this message, keeping the Subject: header intact,
>> Mailman will discard the held message. Do this if the message is
>> spam. If you reply to this message and include an Approved: header
>> with the list password in it, the message will be approved for posting
>> to the list. The Approved: header can also appear in the first line
>> of the body of the reply.
>>
>
>
>
> --
> 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/20131120/1a860bab/attachment-0001.htm

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

Message: 2
Date: Wed, 20 Nov 2013 18:51:27 +0000
From: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Subject: Re: [Moses-support] Moses-support post from
colin.a.cherry@gmail.com requires approval
To: Colin Cherry <colin.a.cherry@gmail.com>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEKMkbiVva+AErX48z4qhuSmH5X=TmX7nULppbwod6qhJTpyow@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

can you send me your ini file and the command you ran. What tuning
algorithm are you using? KBMIRA?

Is the moses.ini file in the new format? The format has been radically
changed over the past year, in case you didnt know. The decoder will still
read the old format but the tuning scripts won't understand the old format


On 20 November 2013 18:47, Colin Cherry <colin.a.cherry@gmail.com> wrote:

> Thanks for getting back so quickly.
>
> For the most part, things are working. For example, I can decode existing
> models with no problems, and the training script that produces phrase
> tables also appears to work. Unfortunately, for parameter tuning, extractor
> appears to not be working. In particular, it is coming back with scores.dat
> files that look like this:
>
> SCORES_TXT_BEGIN_0 0 100 9 BLEU
> 0 0 0 0 0 0 0 0 2
> 0 0 0 0 0 0 0 0 2
> 0 0 0 0 0 0 0 0 2
> 0 0 0 0 0 0 0 0 2
> 0 0 0 0 0 0 0 0 2
>
> This seems to be directly related to the MertBleuScorer test suite, which
> is also failing by returning too many 0's, but that might be a red herring.
> If no one else is observing this issue, then that's useful information. I
> may look into MertBleuScorer's problems anyway, if I get a moment.
>
> -- Colin
>
>
> On Wed, Nov 20, 2013 at 11:47 AM, Hieu Hoang <Hieu.Hoang@ed.ac.uk> wrote:
>
>> hi colin,
>>
>> I think i've fixed everything that needed to be fixed for clang
>> compatibility. The unit tests never worked for me on Mac OSX anyway,
>> something to do with boost static libraries.
>>
>> Your build.log looks like everything compiled, but the unit tests failed,
>> like mine.
>>
>> The few times I've used my mac to train and tune, this was ok.
>>
>>
>>> ---------- Forwarded message ----------
>>> From: Colin Cherry <colin.a.cherry@gmail.com>
>>> To: Moses-support@mit.edu
>>> Cc:
>>> Date: Wed, 20 Nov 2013 09:55:59 -0500
>>> Subject: Trouble building on OS X 10.9 (Mavericks)
>>> Hi all,
>>>
>>> I know Hieu has been working on this, but I'm still having some trouble
>>> getting Moses to build on my Mac after upgrading to Mavericks. In
>>> particular, extractor does not seem to be working properly, and relevant
>>> unit tests in MertBleuScorer are failing.
>>>
>>> I've attached my build.log as suggested by the build process. Please let
>>> me know if any other information would be helpful.
>>>
>>> -- Colin
>>>
>>>
>>> ---------- Forwarded message ----------
>>> From: moses-support-request@mit.edu
>>> To:
>>> Cc:
>>> Date:
>>> Subject: confirm 930e0cc9b7909a1ac82c7494444520fd44d8bf94
>>> If you reply to this message, keeping the Subject: header intact,
>>> Mailman will discard the held message. Do this if the message is
>>> spam. If you reply to this message and include an Approved: header
>>> with the list password in it, the message will be approved for posting
>>> to the list. The Approved: header can also appear in the first line
>>> of the body of the reply.
>>>
>>
>>
>>
>> --
>> Hieu Hoang
>> Research Associate
>> University of Edinburgh
>> http://www.hoang.co.uk/hieu
>>
>>
>


--
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/20131120/0b5c7e0f/attachment-0001.htm

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

Message: 3
Date: Wed, 20 Nov 2013 21:37:35 +0000
From: Philipp Koehn <pkoehn@inf.ed.ac.uk>
Subject: [Moses-support] CALL FOR PAPERS: Workshop on Interactive
Language Learning, Visualization, and Interfaces
To: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAAFADDD4DNJDYNZH+Gqa6yqs0Gwubqa9RwdbrMGB6k3mU3x46w@mail.gmail.com>
Content-Type: text/plain; charset="iso-8859-1"

CALL FOR PAPERS

Workshop on Interactive Language Learning, Visualization, and Interfaces
Venue: ACL 2014 in Baltimore, MD
Date: 27 June 2014http://nlp.stanford.edu/events/illvi2014/illvi2014@gmail.com

DESCRIPTION
People acquire language through social interaction. Computers learn
linguistic models from data, and increasingly, from language-based
exchange with people. How do computational linguistic techniques and
interactive visualizations work in concert to improve linguistic data
processing for humans and computers? How can statistical learning
models be best paired with interactive interfaces? How can the
increasing quantity of linguistic data be better explored and
analyzed? These questions span statistical natural language processing
(NLP), human-computer interaction (HCI), and information visualization
(Vis), three fields with natural connections but infrequent meetings.
Vis and HCI are niches in NLP; Vis and HCI have not fully utilized the
statistical techniques developed in NLP. This workshop aims to
assemble an interdisciplinary community that promotes collaboration
across these fields.


THEMES
Three themes will define this workshop.

1. Active, Online, and Interactive Machine Learning
Statistical machine learning (ML) has yielded tremendous gains in
coverage and robustness for many tasks, but there is a growing sense
that additional error reduction might require a fresh look at the
human role. Presently, human inputs are often restricted to passive
annotation in ML research. However, the fields of ML and HCI are both
developing new techniques---such as active learning,
incremental/online learning, and crowdsourcing---that attempt to
engage people in novel and productive ways.

The first theme of this workshop focuses on advancing interactive
machine learning. How do we jointly solve the learning questions that
have been the domain of NLP and address research topics in HCI such as
managing human workers and increasing the quality of their responses?

2. Language-based user interfaces
NLP techniques have entered mainstream use, but the field currently
focuses more on building and improving systems and less on
understanding how users interact with them in real-world environments.
User interface (UI) design decisions can affect the perceived or
actual performance of a system. For example, while machine translation
(MT) quality improved considerably over the last decade, studies found
that human translators disliked MT output for reasons unrelated to
translation quality. Many existing systems present sentence-level
translations in the absence of relevant context, and disrupt rather
than contribute to a translator's workflow.

The second theme of this workshop focuses on improving people's
ability to engage with language-based UIs and work with linguistic
data. How do we best integrate learning methods, user behavior
understanding, and human- centered design methodology?

3. Text Visualization and Analysis
The quantity and diversity of linguistic corpora is swelling. Recent
work on visualizing text data annotated with linguistic structures
(e.g., syntactic trees, hypergraphs, and sequences) has produced tools
that enable exploration of thematic and recurrence patterns in text.
Visual representations built on the outputs of word-level models
(e.g., sentiment classifiers, topic models, and continuous word
embedding models) now power exploratory analysis of legal documents,
political text, and social media content. Beyond adding analytic
value, interactive visualization can also reduce the upfront effort
needed to set up, configure, and learn a tool, as well as promote
adoption.

The third theme of this workshop centers on improving the utility and
accessibility of text analysis tools. How do we pair appropriate NLP
techniques and visualizations to assist both expert and non-technical
users, who encounter a growing amount of linguistic data in their
professional and everyday lives?


We invite papers on any of the following topics:
* Machine learning with human feedback for NLP
* active learning, online learning with human feedback
* Crowdsourcing for interactive model training
* Text and document visualization
* exploratory text analysis, information retrieval, faceted search
* topic modeling, sentiment analysis, continuous word embedding
* Interactive natural language interfaces
* interactive machine translation
* Linguistic data structure visualization
* sequences, syntactic trees, hypergraphs, hierarchical structures
* Visualizations of short text or temporal patterns
* social media, news, communications
* Visualization and statistical graphing tools for linguistic data
* d3, Vega, Tableau, ggplot


IMPORTANT DATES
The deadline for submissions will be 21 March 2014 (23:59 Samoa Time).
Acceptance decisions will be available on 11 April 2014.
Camera-ready versions due 28 April 2014.

This workshop offers two submission tracks:
1. Long papers
2. Extended abstracts (focused contributions, works in progress,
position papers, system demonstrations)

Long paper submissions should follow the two-column format of ACL 2014
proceedings, and not exceed eight (8) pages including references.
Extended abstract submissions should also follow the two-column format
of ACL 2014 proceedings, and not exceed four (4) pages including
references. Submissions should be in PDF.

Some of the long papers will be selected for oral presentation, while
the others will be presented at a poster session alongside the
extended abstracts. Accepted papers will be included in workshop
proceedings to be published by ACL.


PRELIMINARY SCHEDULE
The program will consist of oral presentations, keynotes by speakers
from each of NLP, HCI, and Vis, and a poster session.


ORGANIZERS
Jason Chuang (UW)
Spence Green (Stanford)
Marti Hearst (UC Berkeley)
Jeffrey Heer (UW)
Philipp Koehn (Edinburgh)

PROGRAM COMMITTEE
Cecilia Aragon (UW)
Chris Callison-Burch (UPenn)
Francisco Casacuberta (Valencia)
Allison Chaney (Princeton)
Christopher Collins (UOIT)
John DeNero (Google)
Marian D?rk (Fachhochschule Potsdam)
Jacob Eisenstein (Georgia Tech)
Jim Herbsleb (CMU)
Robin Hill (Edinburgh)
Eser Kandogan (IBM)
Frank Keller (Edinburgh)
Katie Kuksenok (UW)
Laurens van der Maaten (TU Delft)
Chris Manning (Stanford)
Aditi Muralidharan (UC Berkeley)
Burr Settles (DuoLingo)
John Stasko (Georgia Tech)
Fernanda Vi?gas (Google)
Martin Wattenberg (Google)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20131120/440cb2f3/attachment.htm

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

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


End of Moses-support Digest, Vol 85, Issue 35
*********************************************

Tweet