hi paula,
it'll be useful to us if they are able to do some of those projects. How
they do them is up to you and them.
if they want my help, i'd prefer them to
1. tell me what their idea 1st so I can give them some feedback
2. have a github account. Fork the latest code moses from github (press
the 'fork' button). Commit their code into their fork so i can see it
3. Give me some updates of how they're getting on.
*Out-of-Vocabulary (OOV) Word Handing *- To create a new phrase-table type,
make a copy of
moses/TranslationModel/SkeletonPT.*
Rename the class and follow the example in the file to implement your own
code
*Python Interface - *Don't know. I've never used python*
*
*Analysis of results - *There are many directions where this project can
go. Eg.
1. analysing the output of the decoder
2. good visualisation,
3. analysing the effect of changes to parameters of the decoder.
4. measuring translatability of some sentences
There's been lots of recent prior work, eg.
https://github.com/RobinQrtz/DIMwid
*
*
On 1 October 2013 18:04, Paula Estrella <[email protected]> wrote:
> Dear Mr. Hoang, I'm working in the cs dept of the faculty of mathematics,
> astronomy and physics in Córoda, Arg. and there are some students
> interested in the field of smt and I've seen a page with suggested projects
> to improve moses, are the following topics still of interest to moses? if
> so, how do you normally proceed? do you have any particular
> comments/guidelines/wish-list on how to implement them?
> *Out-of-Vocabulary (OOV) Word Handing
> **Python Interface
> **Analysis of results*
>
> Thanks a lot and best regards,
> Paula Estrella
>
>
>
>
--
Hieu Hoang
Research Associate
University of Edinburgh
http://www.hoang.co.uk/hieu
_______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
