Hello, Thanks for the interest in taking part to the GSOC! :)
If they have time, I think that Gilles Louppe or Arnaud Joly would make good mentors since they are from University of Liège too. Gilles wrote his master thesis on recommender systems so he has good knowledge of matrix factorization problems. Too keep things simple, I would first focus on sequential algorithms and keep the parallel version as a bonus if time permits. Remember that you need to provide us not only with an implementation but also with complete unit tests and documentation. Please start contributing to the project as soon as possible. It is a requirement for eligibility to the GSOC. Mathieu On Mon, Mar 25, 2013 at 10:54 AM, Nicolas Trésegnie <nicolas.treseg...@gmail.com> wrote: > Hi, > > > I’m a graduate student at the University of Liège (Belgium). I got a > bachelor in computer science and now I’m a few months away from the end of > my master in bioinformatics and modeling. I have wanted to work on open > source projects for a long time and I would like to start doing that by > participating in Google Summer of Code 2013. Since I like python and I'm > very interested in machine learning I would like to contribute to this > project. I'm currently using scikit-learn for my master thesis. > > > I'm interested in the idea “Online low rank matrix completion”. I read some > articles on the subject and I think it would be great to first implement the > algorithm presented here (http://bit.ly/10CweSv : “Parallel Stochastic > Gradient Algorithms for Large-Scale Matrix Completion”). It parallelize the > computation of the gradients divinding the data in non overlapping chunks > using what the authors named “cyclic partitioning”. It has been tested on > the Netflix Prize data set and seems fast and efficient. In the first time, > the algorithm could be implemented without parallelization. It could then be > multithreaded using the “cyclic partitioning” and could be adapted to online > learning. I’m also interested to know if the same kind of parallelization > could be applied to the other idea “Online non negative matrix > factorization”. > > > The idea page says that the names of possible mentors is not definitive. Are > these names still going to be modified? I would like to be in touch with the > mentors to have their opinion on the algorithm before spending time writing > a full proposal (maybe they have a better approach in mind!). > > > Best regards, > > > Nicolas Trésegnie > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_d2d_mar > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_mar _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
