00:02 < rcurtin> but then it still links against the wrong thing...

00:02 < rcurtin> I'm not sure

00:02 < anveshi> rcurtin: Oh, I misunderstood it completely

00:02 < rcurtin> you could always just try uninstalling libboost-* from your system since it is too old...

00:02 < anveshi> Was actually looking for some work to do before writing GSoC proposal in order to get better understanding of project.

00:04 < anveshi> can you please suggest something, Which can help understanding tree codebase ?

00:16 < rcurtin> anveshi: you can take a look at open tickets, read the TreeType tutorial document, run tree-based algorithms

00:16 < rcurtin> I have to go to dinner so I can't provide much more advice than that right now

00:16 < rcurtin> there are some open tickets on github relating to the tree code also, you may consider searching through the issues list

00:20 < anveshi> rcurtin: Thanks ! let me check out issues on github

00:22 < awhitesong> Peng_Xu: i guess the you need to uninstall libboostxx-dev files from your system. If you have installed the packaged version of the older boost, the use dpkg -S /usr/include/boost/version.hpp to reconfirm the version and then sudo apt-get autoremove package_name to remove the old files.

00:23 < awhitesong> then reinstall the latest boost. try to install through package manager this time. use this ppa for that. sudo add-apt-repository ppa:boost-latest/ppa

00:24 < Peng_Xu> Yeah, I use apt-get remove libboost-* to uninstall all the boost package

00:24 < awhitesong> let me know if that helped. :)

00:24 < Peng_Xu> Then the error is fixed

00:24 < awhitesong> nice. :)

00:24 < Peng_Xu> Thanks for your patient instructions

00:25 < awhitesong> glad i helped. :)

01:11 < awhitesong> rcurtin: Hey! happy that mlpack got selected this time :) Sorry couldn't continue with the mds code last year in summers as i got through gsoc15 then got involved in some other research projects in ml/deeplrng throughout the year. Shangtong also created the pr for the same and created a basic framework, but a lot of improvements have to made in the code (ball trees, eigen decomposition for Isomap etc with the tests). Was thinking to complete the frame

02:18 < rcurtin> awhitesong: yeah, would love to keep workijg on that MDS implementation

02:18 < rcurtin> just never had the time to look into yours and Shangtong's and figure out how to reconcile them

02:19 < rcurtin> also don't forget to submit a PR to add your name to the list of contributors :)

05:15 < Rufor> Hi, does anyone know if there's somewhere I can find how many students have applied for GSoC for over the previous years and/or their acceptance rate?

05:15 < Rufor> Or perhaps someone has a general idea of this information?

05:18 < Rufor> I'm a student looking to get involved and don't have a great idea of the availability of mentors/positions, and it would help me decide how many applications to send out.

05:28 < rcurtin> Rufor: you can probably find better information in the general #gsoc channel about this

05:28 < rcurtin> as for mlpack, we usually seem to get around 50-80 applications, and we usually end up accepting something like 3-5 projects, maybe more, maybe less --- it depends on the quality of the applications

05:29 < rcurtin> my advice in general would be, if you are going to submit an application, focus on only one project instead of several, which will allow you to put together a higher-quality application for just the one project

05:35 < Rufor> rcurtin: Thanks for the reply, and I did mean to ask specifically about mlpack. If I understand correctly, even if my project is not approved, I could still look for some form of voluntary work/contribution since everything is open source, right?

05:36 < Rufor> It is my intention to focus on a single project/application, but I'm just worried about ending up without something to work on over summer if those I could ask for guidance are already overloaded with students.

05:42 < rcurtin> Rufor: that's right, it is open source, so even if you're not accepted you can still contribute

05:42 < rcurtin> there has been a lot of interest in just about every project for mlpack this year, if you've seen the mailing list

05:43 < rcurtin> that definitely doesn't mean that it's not worth submitting an application, if you feel confident that your proposal is good

05:43 < rcurtin> wait, I guess I misread your message

05:43 < rcurtin> anyway, yeah, I think the mentors here should have at least some spare time still on their hands to help with other contributions

05:44 < Rufor> Okay, thanks a lot for taking the time to answer my questions.

05:44 < praveench> Hi

05:45 < praveench> I am trying to execute the tests of FeedForwardNetworkTest

05:45 < praveench> I have done all the steps

05:46 < praveench> when I write bin/mlpack_test -t FeedForwardNetworkTest

05:46 < praveench> this is the error which is printed : Test setup error: no test cases matching filter

06:20 < rcurtin> praveench: did you read the error?

06:20 < rcurtin> it tells you that you have specified a test suite that does not exist

06:24 < wasiq> hey rcurtin a small question,can 2 people get selected for the same idea?

06:28 < rcurtin> it's possible if there are two outstanding applications

06:56 < wasiq> right :) i shall attempt outstanding.

14:30 < zick1234> Hi, I'm Shikhar Murty from India. I'm a 3rd year EE major at IIT, Delhi with elective concentration in computer science. I am currently doing a year long research in deep learning (applied to NLP). I understand quite a few algorithms from Deep learning, and have implemented CNNs, and neural networks. Would love to write code for FastML

14:39 < zick1234> I would like to contribute to either implementing GoogLeNet or write some deep learning modules for FastML. I have read about GoogLeNet and ZFNet and have a good idea about how these algorithms work. How do i start?

15:12 < kirizaki> Hi!

15:12 < kirizaki> You can start from: http://mlpack.org/involved.html

15:12 < kirizaki> visit github and ask here and via mailing list ;)

15:50 < zick1234> Thanks! @kirizaki

15:52 < zick1234> I am keen on doing a GSoC under FastML. I understand that the applications don't open until next week, but is there something I should be doing right away to increase my chances?

15:59 < wasiq> You could solve some bugs if available for the respective ideas,go through the material and interact with the mentors :)

16:35 < palashahuja> zoq, hi

16:36 < palashahuja> dropconnect doesn't need to be developed for arma:Cube right ?

16:50 < Yvenk> what are the bugs that is related with the alterative colloborative filtering methods

17:13 < zick1234> @zoq, I browsed thru all the tickets related to Deep Learning, but it seems almost all of them are nearing an end.. Could you provide a link to a ticket that is related to Deep learning and hasn't been implemented by anyone? Thanks!

18:45 < mk> Hi, I want to participate in GSoC 2015. I am interested in implementing deep learning modules. I have more than 2 years experience in this field and I have implmented CNN, RNN and RCNN in C++. I also have an idea to discuss. Can anyone guide me through basics?

20:38 < awhitesong> rcurtin: hi. updated the copyright. thanks. also just saw the irc logs, strange the message i sent yesterday truncated and didn't reach completely. so was curious, I guess manifold's been off the ideas this year. Is it because stephentu couldn't(?) be there this year or superLU wasn't done yet?

20:55 < travis-ci> awhitesong/mlpack#14 (copyright - 675101c : Dhawal Arora): The build passed.

20:55 < travis-ci> Change view : https://github.com/awhitesong/mlpack/compare/f0e331889a8a^...675101ceaab8

20:55 < travis-ci> Build details : https://travis-ci.org/awhitesong/mlpack/builds/115571450

21:50 < toshad> hi

21:58 < uzipaz_> hey toshad

21:58 < uzipaz_> what brings you here?

21:58 < toshad> I am interested in contributing in mlpack for GSOC 2016

21:59 < uzipaz_> thats quite nice... do you have any specific topic in mind ?

21:59 < toshad> I am interested in the projects 'Implement tree types' and 'Fast k-centers algorithm'

22:00 < toshad> Could you suggest any pointers to get started?

22:00 < uzipaz_> nice... I am myself looking forward to explore 'parallel stochastic gradient descent' and 'neuroevolution for playing games'

22:01 < uzipaz_> you can go to there github wiki page, there is one article especially written for gsoc 2016 applicants

22:01 < uzipaz_> https://github.com/mlpack/mlpack/wiki/Google-Summer-of-Code-Application-Guide

22:10 < toshad> thanks!

23:23 < uzipaz_> Has anyone compiled and installed mlpack in Win7 using Cygwin?