[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [recommenders-dev] A student interested in GSoC 2013 project idea
|
Hi Rushabh,
> Thanks a lot for your detailed reply. I highly appreciate it. Thanks to
> that I have a much clear idea about the project now.
>
> So as I understand it, a lot depends on the context of the code. There
> are recommendation engines already in place,
yes. At the moment, we propose "instance methods to call" and "methods
to override". However, it doesn't have to end there: students build
prototype of "types to use" and "annotations to use" and Cheng's
recommender in the incubator proposes "arguments to use on a method
call". Another thing we would like to propose are "ways to get an
instance of X".
> the challenge is to
> aggregate the rankings obtained from each so that end result is best
> suitable and most convenient for the programmer.
Yes. Usability is the number 1 goal.
> Just wanted you opinion on this : How about having a model
> (meta-recommender, as you call it) which will tailor/sort the list
> according to the user?
Yes, I can absolutely envision having a meta-recommender that learns
from your interaction with the IDE how to best sort the groups of
recommendations based on the current context.
There doesn't have to be a big ML job running on a server with lots of
data; you can probably refine the meta-recommenders proposals just by
watching which proposal the user selects in a situation and learn from
it on-the-fly. (A student recently build a which "instance methods to
call" recommender in this way.)
> Should I go ahead with the proposal or would you like to have some
> background tasks done from me?
For the meta-recommender thing, it might be worthwhile toying with the
statistics collector Marcel posted a few days ago [1].
Also, you may want to investigate what it would take to tackle this bug
[2]. Please have a look at how subwords matching is currently
implemented and think about how it would interact with multiple
recommenders.
Once you have an understanding of the domain, draft a proposal.
That being said, if you want to tackle a bug first, we won't say "No!" ;-)
Best wishes,
Andreas
[1]
https://docs.google.com/document/d/1oPiTw5ZRnke61eSj3JRVabj0xcDO5rtLuOLJgobsoYY/edit?usp=sharing
[2] https://bugs.eclipse.org/bugs/show_bug.cgi?id=393794
--
Codetrails.com – The knowledge transfer company
