Moses-support Digest, Vol 99, Issue 2

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: equal probabilities (Hieu Hoang)
2. Re: (no subject) (Hieu Hoang)


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

Message: 1
Date: Thu, 01 Jan 2015 21:25:12 +0530
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] equal probabilities
To: Patrick Neurichter <patrickneurichter@gmail.com>,
moses-support@mit.edu
Message-ID: <54A56DE0.7080804@gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed

to post to the moses mailing list, you must subscribe to it 1st. You can
subscribe here
http://mailman.mit.edu/mailman/listinfo/moses-support

Tou answer your question - the translation probabilities is just 1 or
many scores. There's language model scores, word penalty,
phrase-penalty,....

You should take a course on SMT and/or read a book on it

On 29/12/14 14:35, Patrick Neurichter wrote:
> Hello,
>
> could someone please explain how Moses chooses the correct translation
> (in German) if it encounters equal probabilities? Let's say we have a
> sentence "my friend sings a song" (in German = Mein Freund singt ein
> Lied).
>
> my
> *************
> e ..... p(e|f)
> Mein ..... 0.5
> Meine ..... 0.5
>
> friend
> *************
> e ..... p(e|f)
> Freund ..... 0.5
> Freundin ..... 0.5
>
> sings
> *************
> e ..... p(e|f)
> singt ..... 1
>
> a
> *************
> e ..... p(e|f)
> ein ..... 1
>
> song
> *************
> e ..... p(e|f)
> Lied .... 0.5
> Gesang 0.5
>
> So how is Moses going to estimate the translation probabilities and
> choose the correct translation, if there are NO more sentences in a
> parallel corpus that could affect the probabilities?
>
> Pat
>



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

Message: 2
Date: Thu, 01 Jan 2015 21:31:58 +0530
From: Hieu Hoang <hieuhoang@gmail.com>
Subject: Re: [Moses-support] (no subject)
To: Dawit Mulugeta <mulugetadawit@gmail.com>, moses-support@mit.edu
Message-ID: <54A56F76.2060108@gmail.com>
Content-Type: text/plain; charset=windows-1252; format=flowed

are you sure your version of IRSTLM is 5.80.06? Is there a reason you
need to use IRSTLM, as opposed to KENLM, which is built into Moses?

On 01/01/15 14:11, Dawit Mulugeta wrote:
> : ?irstlm? is not a



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

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


End of Moses-support Digest, Vol 99, Issue 2
********************************************

Tweet