00:29 < rcurtin> ShikharJ: yeah, it's all over the map, it depends on the particular workshop

00:29 < rcurtin> many workshops don't actually have formal proceedings either

00:29 < rcurtin> sometimes, the submission is actually just an extended abstract

00:30 < rcurtin> but it is true that in general, NIPS workshop papers are not particularly technically in-depth

00:30 < rcurtin> (the workshops are totally different things than the main conference, usually for incomplete or in-progress type work, where some good results are known but it's not ready for a full submission yet)

00:30 < rcurtin> (so, e.g., submitting a workshop paper lets you get it in front of the relevant audience and maybe have a nice discussion about how to improve it)

00:31 < rcurtin> jeffin143: no worries about the spaces, it's easy for me to fix during merge :)

00:31 < rcurtin> I don't know about webpack, maybe can you send a link or something?

00:59 < rcurtin> Toshal: yeah, I took a look through, I see that it would be a bunch of work, but I'm happy to help if you like. I think that we should put `exactObjective` last, in general, for reverse compatibility

00:59 < rcurtin> there are some situations where UpdatePolicyType parameters are currently last, and in these cases, I agree, maybe it is better to have `exactObjective` before those

01:00 < rcurtin> here's an idea: we could add both constructors, and mark the one that's reverse-compatible as deprecated, and remove it eventually

01:00 < rcurtin> perhaps this would be the best of both worlds? let me know what you think

02:45 < travis-ci> robertohueso/mlpack#44 (mc_kde_error_bounds - a512a48 : Roberto Hueso Gomez): The build was broken.

02:45 < travis-ci> Change view : https://github.com/robertohueso/mlpack/compare/a34193dde720...a512a480a786

02:45 < travis-ci> Build details : https://travis-ci.org/robertohueso/mlpack/builds/561551162

06:04 < jeffin143> could anyone help me debug this code please : https://pastebin.com/wnpk1Xxc

06:04 < jeffin143> not showing complete output

06:05 < jeffin143> rcurtin : https://webpack.js.org/ , They do it for only .js files

06:08 < jeffin143> also python and java by deafult does something like ,for python : from numpy import arange and for java : import static java.lang.Math.abs; But i agree it would be difficult for c++

06:58 < ShikharJ> rcurtin: I see, I was surprised when I saw a 4 page accepted workshop submission in MLOSS, which wasn't even technical. No wonder grad school committees do not look highly upon workshop publications.

08:03 < jenkins-mlpack2> Project docker mlpack nightly build build #393: STILL UNSTABLE in 3 hr 49 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/393/

10:37 < zoq> favre49: Right now there is no good solution for passing a python object/class to mlpack, that's whay there is no binding for the RL code. The same applies to the exectuable, a user can configure the method itself but I don't think there is an easy to way to pass an env. The only option I can see right now, is to let the user choose something what is already there, e.g. CartPole, MountainCar, etc. MEans a

10:37 < zoq> user would have to define the env in C++ first.

10:44 < zoq> favre49: About phased searching, what about we create an artificial task (NEAT is already preconfigured) where we can see that phased searching does make a difference; we don't let it run until it converges, we could just run it once and see if we get the expected output.

10:54 < zoq> jeffin143: Either replace ss.str(line); with stringstream ss(line); or use ss.clear() after the getline while loop: http://www.cplusplus.com/forum/beginner/203521/ does have an explanation.

11:59 < jeffin143> Thansk zoq, That worked out, Thanks for the article too :)

12:00 < jeffin143> Also need an help, https://travis-ci.org/mlpack/mlpack/jobs/561648772 - is there any issue with new branch ?

12:01 < jeffin143> Ann layer test is failing

12:02 < zoq> jeffin143: Yeah, I saw the issue but haven't had the time to take a closer look. You can ignore the issue.

12:02 < jeffin143> Cool, Thanks :)

12:25 < favre49> zoq Alright, thanks. I'll remove neat_main.cpp then, I don't think we can have NEAT bindings.

12:25 < favre49> I'll try what you suggested with the phased searching test

12:44 < zoq> favre49: I don't mind to keep the executable, a user could pass a string that defines the env, if it doesn't exsist we return an error. What do you think?

13:13 < favre49> zoq: wouldn't that mean the user could only use existing envs?

13:13 < favre49> That doesn't seem particularly useful beyond testing purposes

13:20 < rcurtin> ShikharJ: yeah, it is true that graduate school admissions would generally view workshop papers less favorably than conference or journal papers; however, it depends a lot on the workshop, content, etc., and a workshop paper certainly counts for something :)

13:20 < rcurtin> jeffin143: yeah, I see what you mean, thanks for the link. I'm not sure if anything similar exists for C++ though

13:22 < zoq> favre49: It's certainly no ideal, but that would give a user the option to first implement an env in C++ and use the executable afterwards.

13:23 < zoq> favre49: If you like to remove the main for now, that's okay for me as well.

13:24 < rcurtin> jeffin143: also, yeah, looks like one of the ANN layerbtests is failing; there's been an issue open for it for some time if I remember right

15:06 < cerealkiller> Hello everyone!

15:08 < cerealkiller> I'm Nirmesh Bansal, a 2nd year student at The LNM Institute of Information Technology (Communication and Computer Engineering). I would love to contribute here. As of now, I know a bit of C++ and machine learning.

15:46 < favre49> cerealkiller: Hey! mlpack.org/community.html will probably help

15:48 < favre49> You can check out what issues we have open and fix them, or try implementing something new

15:49 < favre49> There's also an optimization library, ensmallen, which you can check out at github.com/mlpack/ensmallen

19:50 < travis-ci> robertohueso/mlpack#45 (mc_kde_error_bounds - fee8261 : Roberto Hueso Gomez): The build is still failing.

19:50 < travis-ci> Change view : https://github.com/robertohueso/mlpack/compare/a512a480a786...fee82619f2b9

19:50 < travis-ci> Build details : https://travis-ci.org/robertohueso/mlpack/builds/561778393