[mlpack] Gsoc2016

Ryan Curtin ryan at ratml.org
Wed Mar 2 11:11:05 EST 2016


On Wed, Mar 02, 2016 at 01:51:45AM +0530, raj goyal wrote:
> Hi everyone,
> 
> I am Rajat goyal 2nd year CS undergraduate . I am interested in doing this
> project
> 1. Implementing various tree types compatible with dual tree algorithms "

Hi Rajat,

There is lots of information available if you look through the mailing
list archives.  The tree types project is a recurring project, so there
has been a lot of discussion about it over the past few years.  Here are
some links to mailing list threads that have more information from
previous years:

https://mailman.cc.gatech.edu/pipermail/mlpack/2014-March/000316.html
https://mailman.cc.gatech.edu/pipermail/mlpack/2014-February/000265.html

Probably one of the most important things to do is read the following
paper and make sure you are familiar with it; trees in mlpack are used
in dual-tree algorithms, so a good understanding of dual-tree algorithms
is pretty necessary.

http://www.ratml.org/pub/pdf/2013tree.pdf

Also, the tree API has changed slightly since the mlpack 2.0.0 release,
and is documented pretty thoroughly here:

http://mlpack.org/docs/mlpack-git/doxygen.php?doc=trees.html

> 2.Neuroevolution algorithms compatible with mlpack.
> as part of gsoc2016 .I have compiled mlpack on Ubuntu and started analyzing
> the source code.I have good working knowledge of c++ and data structres and
> machine learning techniques .I would be glad if anyone can guide me for
> acquiring required knowledge for project .

This response here was for the deep learning modules project, but I
think maybe some of the papers linked to there might be helpful to read
and be familiar with:

https://mailman.cc.gatech.edu/pipermail/mlpack/2016-February/000738.html

Definitely reading papers on, e.g., CNE, CMA-ES, NEAT, or HyperNEAT
would be a good place to start.  There  is a nice reading list on the
Ideas page:

https://github.com/mlpack/mlpack/wiki/SummerOfCodeIdeas#neuroevolution-algorithms

Let me know if I can clarify anything.

Thanks!

Ryan

-- 
Ryan Curtin    | "I'm just going to shoot you once!"
ryan at ratml.org |   - Joseph Dunn



More information about the mlpack mailing list