Re: [DL-Learner discussion] Adding DL Learner Components (specifically, Refinement Operators)

[Jens L. <le...@in...>]

Hello Matthew,

sorry for the very late reply.

Am 29.06.2012 18:37, schrieb Matthew Byrne:
> Hi All
>
> I'm trying to experiment with modifications to DL-Learner initially by
> experimenting with different Refinement Operators. To get the ball
> rolling, I copied RhoDRDown.java and:
> - changed the name to MpbVersion1 and adjusted the annotation to "mpb"
> (where previously there was "rho")
> - included a reference to my new file in the "componentClassNames" list
> of AnnComponentManager.java
> - then tried to run it by changing rho to mpb in an already-working con
> file.
>
> However it does not work.  Any ideas how I can make it work?

In general what you did is correct. However, not all DL-Learner 
components are actually implemented in a flexible way. You got the 
exception, because the used learning algorithm only accepted the 
RhoDRDown refinement operator instead of using an interface. I just 
spend some time to abstract away from this for the CELOE learning 
algorithm, so your problem should now be solved.

I added some documentation here during implementation:
http://dl-learner.org/wiki/WritingComponents

We appreciate further feedback and also source code contributions.

Sorry again for the delay. Usually, we are faster in replying mails.

Kind regards,

Jens

-- 
Dr. Jens Lehmann
Head of AKSW/MOLE group, University of Leipzig
Homepage: http://www.jens-lehmann.org
GPG Key: http://jens-lehmann.org/jens_lehmann.asc