Moses-support Digest, Vol 99, Issue 19

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. [decoding-graph-backoff] (Hieu Hoang)
2. Re: [decoding-graph-backoff] (Marcin Junczys-Dowmunt)
3. Re: [decoding-graph-backoff] (Hieu Hoang)
4. Re: [decoding-graph-backoff] (Raj Dabre)


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

Message: 1
Date: Fri, 9 Jan 2015 15:08:21 +0000
From: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Subject: [Moses-support] [decoding-graph-backoff]
To: moses-support <moses-support@mit.edu>
Message-ID:
<CAEKMkbifApkyVMecOO--NWijpi8Em0Q6VnoFoftZjXUnhctiZQ@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hi All

Does anyone use this functionality in Moses when you have multiple
phrase-tables?

>From the code, it doesn't look like it works as described in
http://www.statmt.org/moses/?n=Moses.AdvancedFeatures

Maybe I'm missing something

--
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/20150109/b199c36b/attachment-0001.htm

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

Message: 2
Date: Fri, 09 Jan 2015 16:11:13 +0100
From: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Subject: Re: [Moses-support] [decoding-graph-backoff]
To: moses-support@mit.edu
Message-ID: <54AFEF91.2010605@amu.edu.pl>
Content-Type: text/plain; charset="windows-1252"

Hm, we have been using it at WIPO, but I have to admit I never checked
it _actually_ does anything useful. We sorta believe it does.

W dniu 09.01.2015 o 16:08, Hieu Hoang pisze:
> Hi All
>
> Does anyone use this functionality in Moses when you have multiple
> phrase-tables?
>
> From the code, it doesn't look like it works as described in
> http://www.statmt.org/moses/?n=Moses.AdvancedFeatures
>
> Maybe I'm missing something
>
> --
> Hieu Hoang
> Research Associate
> University of Edinburgh
> http://www.hoang.co.uk/hieu
>
>
>
> _______________________________________________
> 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/20150109/631f4e82/attachment-0001.htm

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

Message: 3
Date: Fri, 9 Jan 2015 15:20:44 +0000
From: Hieu Hoang <Hieu.Hoang@ed.ac.uk>
Subject: Re: [Moses-support] [decoding-graph-backoff]
To: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Cc: moses-support <moses-support@mit.edu>
Message-ID:
<CAEKMkbimdNrWkaaqhWtgV4_AccWDEvkg0iP18FYBDPt+7MfUdA@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

>From the git history, I think it was slightly broken, then I broke it even
more in May 2014.

https://github.com/moses-smt/mosesdecoder/commit/44fec57c535db2df73ccbb1628d8143a9c728c19

It was
endPos-startPos+1 >= backoff
then
endPos-startPos+1 <= backoff
I think it should be
endPos-startPos+1 > backoff

I'll change it if it's ok with everyone


On 9 January 2015 at 15:11, Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
wrote:

> Hm, we have been using it at WIPO, but I have to admit I never checked
> it _actually_ does anything useful. We sorta believe it does.
>
> W dniu 09.01.2015 o 16:08, Hieu Hoang pisze:
>
> Hi All
>
> Does anyone use this functionality in Moses when you have multiple
> phrase-tables?
>
> From the code, it doesn't look like it works as described in
> http://www.statmt.org/moses/?n=Moses.AdvancedFeatures
>
> Maybe I'm missing something
>
> --
> Hieu Hoang
> Research Associate
> University of Edinburgh
> http://www.hoang.co.uk/hieu
>
>
>
> _______________________________________________
> Moses-support mailing listMoses-support@mit.eduhttp://mailman.mit.edu/mailman/listinfo/moses-support
>
>
>
> _______________________________________________
> 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/20150109/9781746f/attachment-0001.htm

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

Message: 4
Date: Sat, 10 Jan 2015 00:24:02 +0900
From: Raj Dabre <prajdabre@gmail.com>
Subject: Re: [Moses-support] [decoding-graph-backoff]
To: Marcin Junczys-Dowmunt <junczys@amu.edu.pl>
Cc: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAB3gfjC+7Thqr3enJJKSf6=LsBdLg+ABiQceezBFqpmv7ZjuAw@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hey,

I have used the multiple decoding paths with and without the backoff
procedures.
I once experimented with using a French-English phrase table as a backoff
table when translating from Creole (Mauritian) to English (I had a small
Creole-English corpus).
I found that I got a decent improvement in BLEU when I used a 4 gram
backoff setting.

In my opinion one should use the multiple Decoding Paths when confident
that the 2nd (and other) table has decent quality. (Domain adaptation
relies on this AFAIK)
But if the 2nd table fulfills the role of a supplementary table (as french
and creole have common words the French-English can provide OOV support)
then backoff is better.

Regards.

On Sat, Jan 10, 2015 at 12:11 AM, Marcin Junczys-Dowmunt <junczys@amu.edu.pl
> wrote:

> Hm, we have been using it at WIPO, but I have to admit I never checked
> it _actually_ does anything useful. We sorta believe it does.
>
> W dniu 09.01.2015 o 16:08, Hieu Hoang pisze:
>
> Hi All
>
> Does anyone use this functionality in Moses when you have multiple
> phrase-tables?
>
> From the code, it doesn't look like it works as described in
> http://www.statmt.org/moses/?n=Moses.AdvancedFeatures
>
> Maybe I'm missing something
>
> --
> Hieu Hoang
> Research Associate
> University of Edinburgh
> http://www.hoang.co.uk/hieu
>
>
>
> _______________________________________________
> Moses-support mailing listMoses-support@mit.eduhttp://mailman.mit.edu/mailman/listinfo/moses-support
>
>
>
> _______________________________________________
> Moses-support mailing list
> Moses-support@mit.edu
> http://mailman.mit.edu/mailman/listinfo/moses-support
>
>


--
Raj Dabre.
Research Student,
Graduate School of Informatics,
Kyoto University.
CSE MTech, IITB., 2011-2014
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20150110/2f03125d/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 19
*********************************************

Tweet