opennlp-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rodrigo Agerri <rage...@apache.org>
Subject Re: WSD-Supervised
Date Tue, 11 Aug 2015 08:28:54 GMT
Hi Mondher,

Yes, the preprocessing changes are all the same. With respect to the
CLI and the Evaluator, they are common regardless of the type of WSD
(unsupervised or supervised) so I was assuming that you were
coordinating with Anthony as you have done before.

Please note that it is very important to have these issues completed
before the GSOC (including the all words algorithm). Also note that
GSOC students are supposed to be working 40 hours a week on their
projects.

Best,

Rodrigo

On Mon, Aug 10, 2015 at 8:23 PM, Mondher Bouazizi
<mondher.bouazizi@gmail.com> wrote:
> Dear Rodrigo,
>
> I am sorry I for my late reply. I am not in Japan these days. I have been to
> a conference, then I went to Tunisia. I am currently working on the all
> words disambiguation. Hopefully it will be done in the next few days. I need
> first to take into consideration the changes Anthony made in the code. He
> told me that J├Ârn asked him to remove as much preprocessing as possible, so
> I guess I have to do the same.
>
> Regarding the evaluator, it has already been implemented. Please refer to
> the class "IMSEvaluatorTest" to see the performances of the IMS
> disambiguator on the senseval data set.
>
> Yours sincerely,
>
> Mondher
>
> On Monday, August 10, 2015, Rodrigo Agerri <agerri.rodrigo@gmail.com> wrote:
>>
>> Hello Mondher,
>>
>> How is the all words IMS disambiguation progressing? We really need to
>> focus on this to have a good candidate for integration in opennlp
>> tools. The evaluator, the CLI and the all words supervised
>> disambiguation should be the focus.
>>
>> Cheers,
>>
>> Rodrigo
>>
>> On Sat, Jul 18, 2015 at 5:40 AM, Mondher Bouazizi
>> <mondher.bouazizi@gmail.com> wrote:
>> > Dear Rodrigo,
>> >
>> > I have made some major modifications on my part of the WSD component. I
>> > attached the patch to [1]. The main modifications are as follows:
>> >
>> >    - Fixed the IMS approach (Supports now with Semsor3.0 data).
>> >    - Implemented the IMS Evaluator.
>> >    - Added and clarified some parts of the documentation.
>> >    - Added "Semeval3" and "Semcor3.0" readers.
>> >
>> > Please find the patch and more details about modification on the
>> > following
>> > link [1].
>> >
>> > Mondher
>> >
>> >
>> > [1] https://issues.apache.org/jira/browse/OPENNLP-790

Mime
View raw message