03:30 < KimSangYeon-DGU> Hi zoq! I thought I can work on `gym_tcp_api` maintenance work for compatibility with OpenAI gym's latest version, but I can't find time to do. Actually, I'm interested in `Quantum Gaussian Mixture Models` project for GSoC 2019. I think I should spend time in doing the project and related works like #1666. So, I posted new issue about the maintenance work for `gym_tcp_api`.

03:36 < KimSangYeon-DGU> zoq: I'm sorry...

06:16 < Me-so-cool> JOIN

06:19 < Me-so-cool> Hello guys! I'm really new to open source. I have experience in competitive programming though.

06:20 < Me-so-cool> How and where should I start contributing to mlpack?

06:20 < Me-so-cool> I see many people contributing are of high level like PhD and others I'm merely a sophomore

06:20 < Me-so-cool> How can I help in any way possible? Thanks!

06:24 < KimSangYeon-DGU> Me-so-cool: Hi, I think the link is helpful to you. http://www.mlpack.org/involved.html

06:26 < Me-so-cool> Yeah I have been there. I setup the environment and all

06:26 < Me-so-cool> But I don't know where to start. Like the package is so big and complex and I don't understand most of the part

06:27 < Me-so-cool> Take linear regression for example. It's already implemented in the best way possible

06:27 < Me-so-cool> What could a beginner like me do to improve it. Thanks and sorry if I'm being dumb

06:27 < KimSangYeon-DGU> How about the issue page in Github? There are many interesting work to do.

06:28 < KimSangYeon-DGU> *works

06:29 < Me-so-cool> Okay sure will try.

06:30 < Me-so-cool> By the way I'm thinking of taking part in gsoc this summer. Are beginners like me eligible?

06:32 < KimSangYeon-DGU> Ah, actually, I'm not a member of mlpack yet. so I'm not qualified to answer the question.

06:34 < Me-so-cool> Okay cool thanks

09:01 < jenkins-mlpack2> Project docker mlpack nightly build build #245: STILL UNSTABLE in 3 hr 47 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/245/

13:07 < sreenik> ShikharJ: Sorry I wasn't active for a while. I wanted to ask if there was some predefined module implementing commonly used data augmentation techniques like selective cropping, rotation, etc.

15:10 < rcurtin> looks like gitdub is going crazy again with the emails, I'm not sure why

15:31 < ShikharJ> rcurtin: Did you happen to take a look at the testing wiki? I think we should mention it in the bot's initial message.

15:39 < rcurtin> I haven't had a chance; let me add that to my todo list for today

15:43 < jenkins-mlpack2> Project mlpack - git commit test build #115: STILL UNSTABLE in 43 min: http://ci.mlpack.org/job/mlpack%20-%20git%20commit%20test/115/

15:43 < jenkins-mlpack2> noreply: Merge pull request #1738 from KimSangYeon-DGU/compiler

15:52 < Akfluffy> Hey guys, do I have to encode input data to be between -1 and 1? Or is it done automatically ?

16:12 < ShikharJ> rcurtin: Ah okay. No hurry!

16:18 < rcurtin> ShikharJ: just took a look, really nice :) I do think we should add a section on test cases that depend on random output, which pass Travis's single check but fail when they get to Jenkins

16:18 < rcurtin> the right way to test random tests like that is to set the random seed and run many times, like in the description of the bug I just filed:

16:18 < rcurtin> https://github.com/mlpack/mlpack/issues/1740

16:20 < ShikharJ> rcurtin: Agreed, a nice idea. I'll cover that as well. As for the test writing using Boost, and reporting and analyzing failures, I was hoping you or Marcus would be able to better express the methodology there, so I left it. I'll make the changes soon.

16:27 < jenkins-mlpack2> Project mlpack - git commit test build #116: STILL UNSTABLE in 44 min: http://ci.mlpack.org/job/mlpack%20-%20git%20commit%20test/116/

16:27 < jenkins-mlpack2> noreply: Merge pull request #1726 from mirraaj/extend1679

16:37 < rcurtin> ShikharJ: we can just link to the Boost Unit Test Framework documentation and maybe to an existing test somewhere

16:37 < rcurtin> I think the Boost tests are pretty easy to imitate

16:39 < ShikharJ> rcurtin: Hmm, that should work out fine as well. Thanks for the ideas :)

16:39 < rcurtin> sure, happy to help