[mlpack] Hello Everyone

[Ryan Curtin]

On Wed, Jan 25, 2017 at 07:03:39PM +0530, Vivek Pal wrote:
> I am Vivek Pal, a senior year undergraduate student with major in Computer
> Science at Delhi Technological University in New Delhi, India. I'm looking
> forward to be a part of mlpack in GSoC 2017 and beyond. I'm just getting
> started afresh with mlpack, although I've used it for some hackish ML
> experiments in the past.
> 
> I do have some experience with ML through some academic projects that I've
> done over the past 1-2 years ranging from building a recommender system to
> working on sentiment analysis of tweets etc. and thus, I have developed
> strong interests in ML. However, I hope being an ML practitioner or expert
> is not a primary requirement for a GSoC student in mlpack. :)
> 
> Also, last semester I took a course that was entirely based on NNs and it
> got me really interested into this subset of machine learning. I was able
> to publish an article based on single layer neural net intended for
> beginners here: http://www.geeksforgeeks.org/implementing-ann-training-
> process-in-python/
> 
> Since this year's project ideas are not out yet, I have looked up previous
> year ideas and find two interesting and feel confident enough to work on
> (if they're planned to be continued this year):
> 
> 1. Alternatives to neighborhood-based collaborative filtering
> 2. Essential Deep Learning Modules
> 
> I'd appreciate advice based on these project ideas to help me explore more
> about these and be able to finally make sound decision of choosing between
> two.
> 
> I've already started to familiarise myself with the related codebase and
> will find some warm up tasks right away from issues list. However, it'd be
> great if I could get some pointers for going ahead that would help me
> understand codebase and tests better.
> 
> I should probably also mention that I was a part of GSoC last year. I
> worked with Xapian which is a probabilistic search engine library written
> in C++. Broadly, my work was focused on improving the existing weighting
> schemes in Xapian which basically started with reading a bunch of research
> papers. I was able to successfully complete all the set out goals of the
> project and the work product was merged in the Xapian codebase.

Hi Vivek,

Welcome to the community.  Being a machine learning expert is certainly
helpful for a GSoC student but what's sometimes more important is a good
background in C++ and software development along with a willingness to
learn.

I'm still willing to mentor the CF idea.  If you search the mailing list
archives, I seem to remember there being a lot of discussion about the
idea last year, and maybe you can find some insights there.  In the end,
anything that helps the CF code get easier to use or more functional so
that users can easily build a recommendation system would be a great
project.

I also remember a lot of discussion about the deep learning modules
project; I'm sure you can find a lot about that one in the archives too.

Thanks,

Ryan

-- 
Ryan Curtin    | "Kill them, Machine... kill them all."
ryan at ratml.org |   - Dino Velvet