Moses-support Digest, Vol 110, Issue 43

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: Doubts on Multiple Decoding Paths (Michael Denkowski)
2. Using multiple translation model (Paper) (Cyrine NASRI)
3. Strange behavioure of the decoder (Sehrob Ibrohimov)


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

Message: 1
Date: Fri, 25 Dec 2015 13:11:53 -0500
From: Michael Denkowski <michael.j.denkowski@gmail.com>
Subject: Re: [Moses-support] Doubts on Multiple Decoding Paths
To: Anoop (?????) <anoop.kunchukuttan@gmail.com>
Cc: moses-support <moses-support@mit.edu>, Philipp Koehn <phi@jhu.edu>
Message-ID:
<CA+-GegKTnPUZsogwuD5Oc=Df+wMNr47YS0Y4avfmyzrYArWKwA@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hi Anoop,

The Moses documentation for PhraseDictionaryGroup
<http://www.statmt.org/moses/?n=Advanced.Models#ntoc7> that you originally
referenced is the most up-to-date description of this capability in Moses.
If you're looking for something to cite, the original version of the
"union" capability in Moses (arbitrary number of models, full feature set
from each, and zeros when a phrase is not in a model) is described in my
thesis "Machine Translation for Human Translators
<https://www.cs.cmu.edu/~mdenkows/pdf/denkowski-mtht-2015.pdf>", Sections
3.4 and 3.4.1. There it's being used for a very specific application so
it's described in that context rather than as a general model combination
approach. If anything still isn't clear, it probably means I should expand
the Moses documentation, so let me know. : )

Best,
Michael

On Fri, Dec 25, 2015 at 3:40 AM, Anoop (?????) <anoop.kunchukuttan@gmail.com
> wrote:

> Hi Michael,
>
> Is there a publication/document describing the *union *option or its use
> in an MT scenario?
>
> Regards,
> Anoop.
>
> On Sat, Dec 19, 2015 at 3:38 AM, Michael Denkowski <
> michael.j.denkowski@gmail.com> wrote:
>
>> Hi Anoop,
>>
>> Confirming that your reading of "union" is in fact how it works. If you
>> want each phrase to be scored by all tables without having to worry about
>> making sure every phrase is in every table, I you can use
>> PhraseDictionaryGroup with default-average-others=true. This multiplies
>> the size of the phrase feature set by the number of models, so I recommend
>> running mer-moses.pl with --batch-mira.
>>
>> Best,
>> Michael
>>
>> On Fri, Dec 18, 2015 at 1:08 PM, Philipp Koehn <phi@jhu.edu> wrote:
>>
>>> Hi,
>>>
>>> that sounds right.
>>>
>>> The "union" option is fairly new, developed by Michael Denkowski.
>>> I am not aware of any empirical study of the different methods,
>>> so I'd be curious to see what you find.
>>>
>>> -phi
>>>
>>> On Fri, Dec 18, 2015 at 1:35 AM, Anoop (?????) <
>>> anoop.kunchukuttan@gmail.com> wrote:
>>>
>>>> Hi,
>>>>
>>>> I am trying to understand the multiple decoding paths feature in Moses.
>>>>
>>>> The documentation (http://www.statmt.org/moses/?n=Advanced.Models#ntoc7)
>>>> describes 3 methods: both, either and union
>>>>
>>>> The following is my understanding of the options. Please let me know if
>>>> it is correct:
>>>>
>>>>
>>>> - With *both* option, the constituent phrases of the target
>>>> hypothesis come from both tables (since they are shared) and are scored
>>>> with both the tables.
>>>> - With *either* option, all the constituent phrases of a target
>>>> hypothesis come from a single table, but different hypothesis can use
>>>> different tables. Each hypothesis is scored using one table only. I did not
>>>> understand the " additional options are collected from the other tables"
>>>> bit in the documentation.
>>>> - With *union* option, the constituent phrases of a target
>>>> hypothesis come from different tables and are scored using scores from all
>>>> the tables. Use 0 if the option doesn't appear in some table, unless the
>>>> *default-average-others=true* option is used.
>>>>
>>>>
>>>> Regards,
>>>> Anoop.
>>>>
>>>> --
>>>> I claim to be a simple individual liable to err like any other fellow
>>>> mortal. I own, however, that I have humility enough to confess my errors
>>>> and to retrace my steps.
>>>>
>>>> http://flightsofthought.blogspot.com
>>>>
>>>> _______________________________________________
>>>> 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
>>>
>>>
>>
>
>
> --
> I claim to be a simple individual liable to err like any other fellow
> mortal. I own, however, that I have humility enough to confess my errors
> and to retrace my steps.
>
> http://flightsofthought.blogspot.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20151225/b21ac763/attachment-0001.html

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

Message: 2
Date: Sat, 26 Dec 2015 11:34:36 +0100
From: Cyrine NASRI <cyrine.nasri@univ-lorraine.fr>
Subject: [Moses-support] Using multiple translation model (Paper)
To: "moses-support@mit.edu" <moses-support@mit.edu>
Message-ID:
<CAPg_V0jpJj90sHsCfTx3KNG0zMN=HYJ_oJUGKxJzteTa-vUGTg@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

Hello,

I would like to know if there is a paper which describe how Moses do to
select the best translation when We use 2 differents translation tables.

Best regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20151226/beb37b55/attachment-0001.html

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

Message: 3
Date: Sat, 26 Dec 2015 18:53:01 +0500
From: Sehrob Ibrohimov <isehrob@gmail.com>
Subject: [Moses-support] Strange behavioure of the decoder
To: moses-support <moses-support@mit.edu>
Message-ID:
<CADJquiA8rr2wuzKsLHS3EYvofX8vMYDfyKyV8sFJEfpeb_+rGw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8

Hello everyone!

I have came accross some strange behaviour of moses decoder and can't
get it right. I can't figure out what the problem is.
The problem is, I have trained translation models about a year ago
with moses (version 3.0) and got pretty good results. The translation
was pretty impressive. But some days ago I have got new machine
(Centos 7) and build moses sources and give him translation model I
had. And I can't get any translations. But these models was giving
good results with the previous moses which was on ubuntu 14.0.
I give him input like: "????????? ?????????? ???????? ???? ?
?????????? ????????? ??????".
It gives me: ?????????|UNK|UNK|UNK ??????????|UNK|UNK|UNK
????????|UNK|UNK|UNK ????|UNK|UNK|UNK ?|UNK|UNK|UNK
??????????|UNK|UNK|UNK ?????????|UNK|UNK|UNK ??????|UNK|UNK|UNK
[11111111] [total=-798.292]
core=(-800.000,-8.000,8.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,-25.136)
Then I thought that there may be some new changes in moses, and I have
retrained my models with the data I had, using EMS but nothing has
changed, the decoder gives me the same results.
Please, help me to figure out what the problem is!

Sincerely yours
Sehrob



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

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


End of Moses-support Digest, Vol 110, Issue 43
**********************************************

Tweet