--- Log opened Sun Jun 10 00:00:43 2018 | ||
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4305 synchronized by shubham808 | 09:58 |
---|---|---|
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4336 opened by vinx13 | 12:51 |
-!- HeikoS [~heiko@host86-151-45-5.range86-151.btcentralplus.com] has joined #shogun | 17:22 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 17:22 | |
@HeikoS | rcurtin: hi | 17:25 |
@HeikoS | wiking: hi | 17:25 |
@HeikoS | rcurtin: wiking I need a bio from you (short), see the MLOSS notes doc | 17:29 |
@HeikoS | bottom | 17:29 |
rcurtin | rcurtin: sure, hang on | 17:37 |
rcurtin | how serious do you want it to be...? ;) | 17:37 |
rcurtin | I guess if it needs to be accepted by the NIPS committee it should be serious | 17:37 |
@HeikoS | just a few sentences that tell us that you are an MLOSS expert | 17:39 |
@HeikoS | rcurtin: thx for editing, think I am kinda done | 17:40 |
@HeikoS | all we would need is a tentative website | 17:40 |
rcurtin | sure, writing now | 17:40 |
@HeikoS | btw I was suggested to inivite the founder of openml | 17:41 |
@HeikoS | so did that | 17:41 |
@HeikoS | if he says no I will write to new ceo of github | 17:41 |
@HeikoS | but I thought openml fits nips better | 17:41 |
@HeikoS | as it is ML + OSS rather than github just being OSS | 17:42 |
rcurtin | I could agree with that, but if you get the new CEO of github to talk anywhere at NIPS, the room will be completely packed | 17:42 |
@HeikoS | deadline is tomorrow | 17:42 |
@HeikoS | do we want that? :D | 17:42 |
@HeikoS | I was there when CEO of facebook came to the deep learning workshop in 2012 | 17:42 |
@HeikoS | unbearable ;0 | 17:42 |
@HeikoS | Well let's see | 17:42 |
rcurtin | that's true | 17:42 |
@HeikoS | I will defo ask if the openml guy says no | 17:42 |
@HeikoS | rcurtin: merci for editing | 17:44 |
@HeikoS | will leave it for now ... | 17:44 |
rcurtin | sure, sounds good | 17:44 |
rcurtin | should be done in 5 minutes or so | 17:44 |
@HeikoS | rcurtin: ah one more thing | 17:48 |
@HeikoS | you said that the gsoc team might fund such things? | 17:48 |
@HeikoS | or was that viktor saying that? | 17:48 |
rcurtin | I had said that; if you like, I can get in touch with Cat Allman | 17:49 |
@HeikoS | yes would you mind? | 17:49 |
rcurtin | I sat through a session she was putting on last year where basically what she had said was, her job is to have Google give various grants to open source projects and initiatives | 17:49 |
@HeikoS | we need travel funding for another person, or two. Plus maybe something more for dinner | 17:49 |
rcurtin | and requests for a few thousand dollars is not too much for most things | 17:49 |
@HeikoS | so if we got like 3000 that would be good | 17:49 |
@HeikoS | yeah pls do get in touch | 17:49 |
@HeikoS | could you cc me? | 17:49 |
rcurtin | she also said to me in person, if we get a paper in based on a GSoC project, she would likely be happy to give more than the $500 that Google already mentions they will give | 17:49 |
rcurtin | absolutely, I will do that next | 17:50 |
@HeikoS | thx | 17:50 |
@HeikoS | wuill you try to get a paper based on gsoc project? pretty cool that would be | 17:50 |
rcurtin | usually we do, but they tend to be simple workshop papers, never anything crazy like ICML or NIPS | 17:50 |
rcurtin | this year we have a student working on an implementation of LMNN | 17:51 |
@HeikoS | ah cool | 17:51 |
@HeikoS | we can do comparisons :) | 17:51 |
rcurtin | and I think there is enough there that we may actually be able to publish some neat results. I believe our implementation will be able to scale to hundreds of thousands of points or more, but there is still some optimization to do :) | 17:51 |
@HeikoS | with fernados impl | 17:51 |
@HeikoS | cool stuff | 17:51 |
rcurtin | yeah, definitely | 17:51 |
@HeikoS | let me know how it goes! | 17:51 |
rcurtin | in his proposal is to write benchmarking scripts for it, so we will be able to see | 17:51 |
rcurtin | HeikoS: how much should I ask for? a few thousand? I know that Cat will want to see a number | 17:59 |
@HeikoS | travel for 2 is 2000 | 18:01 |
@HeikoS | and then maybe a bit for a dinner or so | 18:01 |
@HeikoS | 2.5? | 18:01 |
rcurtin | ok, I'll go ahead and ask for $3k then | 18:01 |
rcurtin | much better to get too much than to need to come back :) | 18:02 |
rcurtin | it's also possible (and I think likely) that she'll say, 'well, we can do half that but not the whole thing...' | 18:02 |
rcurtin | ok, sent an email to cat. now I will get back to making this shelf :) | 18:07 |
-!- witness_ [uid10044@gateway/web/irccloud.com/x-xccspnlusqofaqkj] has joined #shogun | 19:07 | |
-!- HeikoS [~heiko@host86-151-45-5.range86-151.btcentralplus.com] has quit [Quit: Leaving.] | 19:33 | |
-!- HeikoS [~heiko@host86-151-45-5.range86-151.btcentralplus.com] has joined #shogun | 22:29 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 22:29 | |
--- Log closed Mon Jun 11 00:00:45 2018 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!