--- Log opened Mon Apr 08 00:00:26 2013 | ||
-!- foulwall [~foulwall@2001:da8:215:c252:a186:6bb3:d964:b177] has joined #shogun | 03:33 | |
-!- foulwall [~foulwall@2001:da8:215:c252:a186:6bb3:d964:b177] has quit [Ping timeout: 245 seconds] | 03:38 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 04:22 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 246 seconds] | 04:54 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 04:54 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 264 seconds] | 05:12 | |
-!- blackburn [~blackburn@109.226.90.103] has quit [Quit: Leaving.] | 06:21 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 248 seconds] | 06:48 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 06:52 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 07:03 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 256 seconds] | 07:11 | |
-!- sumit_ [75e02ae4@gateway/web/freenode/ip.117.224.42.228] has joined #shogun | 07:27 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has joined #shogun | 08:04 | |
-!- hoijui [~hoijui@dslb-088-075-046-171.pools.arcor-ip.net] has joined #shogun | 08:19 | |
-!- sumit_ [75e02ae4@gateway/web/freenode/ip.117.224.42.228] has quit [Ping timeout: 245 seconds] | 08:41 | |
-!- sonne|work [~sonnenbu@sams-office-nat.tomtomgroup.com] has joined #shogun | 08:49 | |
n4nd0 | http://www.smbc-comics.com/index.php?db=comics&id=2934#comic :D | 09:50 |
---|---|---|
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun | 10:19 | |
blackburn | sonne|work: wiking: n4nd0: 10 hours left! | 10:23 |
n4nd0 | blackburn: yeah! | 10:23 |
n4nd0 | let's keep ourselves crossed | 10:23 |
n4nd0 | not just our fingers | 10:23 |
blackburn | I am not that flexible! | 10:24 |
sonne|work | I hope that doesn't affect # toilet visits | 10:27 |
sonne|work | blackburn: Any idea how one can modify the registration form for the workshop? | 10:28 |
sonne|work | more specifically | 10:28 |
sonne|work | the choices ml expert / bloody beginner | 10:28 |
sonne|work | etc | 10:28 |
blackburn | sonne|work: what do you mean? | 10:29 |
blackburn | ah | 10:29 |
blackburn | Form - Edit form | 10:29 |
blackburn | as usual | 10:29 |
sonne|work | ahh ok - because I have a parse error for Acknowledged with some concepts of machine learning | 10:30 |
sonne|work | I guess - this means machine learning expert? | 10:31 |
sonne|work | ^ blackburn or what? | 10:31 |
-!- hoijui [~hoijui@dslb-088-075-046-171.pools.arcor-ip.net] has quit [Read error: Connection reset by peer] | 10:34 | |
blackburn | sonne|work: your LL(1) gets stucked too often! | 10:34 |
blackburn | jsut fix it! | 10:34 |
sonne|work | blackburn: yeah, I need some context sensitive grammar | 10:35 |
sonne|work | blackburn: I don't know how to properly fix it b'cause I don't know what you meant | 10:35 |
sonne|work | anyway I am writing sth in there | 10:35 |
sonne|work | done | 10:36 |
blackburn | sonne|work: your guess is correct | 10:37 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 10:40 | |
sonne|work | blackburn: soulmate! | 10:40 |
blackburn | sonne|work: you've got to learn runglish | 10:43 |
blackburn | :D | 10:43 |
sonne|work | sorry I only speed denglish | 10:44 |
sonne|work | speak | 10:44 |
blackburn | no you have learnt the basics of runglish | 10:44 |
blackburn | otherwise you weren't able to understand my alphanumerical sequencies | 10:44 |
sonne|work | blackburn: must be due to the roots of these languages being ENGLISH! | 10:56 |
blackburn | sonne|work: no my language origins from mongolian warriors! | 10:57 |
sonne|work | and my origins is Teutonian! It was my grand-grand-grand-grand-grand father who stopped Cesar in the this battle http://en.wikipedia.org/wiki/Battle_of_Teutoburg_Forest | 10:59 |
blackburn | sonne|work: not grand^n enough! | 11:01 |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 11:16 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has quit [Quit: leaving] | 11:31 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 12:14 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun | 12:30 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 12:46 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 245 seconds] | 12:54 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has joined #shogun | 13:31 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has quit [Ping timeout: 240 seconds] | 14:04 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 14:24 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has joined #shogun | 14:28 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Ping timeout: 245 seconds] | 14:29 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has joined #shogun | 14:43 | |
-!- foulwall [~foulwall@li379-21.members.linode.com] has quit [Remote host closed the connection] | 15:05 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 16:39 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has joined #shogun | 17:00 | |
-!- foulwall [~foulwall@2001:da8:215:c252:3969:1a7a:9189:1ce8] has joined #shogun | 17:14 | |
-!- blackburn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 17:26 | |
-!- foulwall [~foulwall@2001:da8:215:c252:3969:1a7a:9189:1ce8] has quit [Ping timeout: 245 seconds] | 17:40 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has quit [Ping timeout: 246 seconds] | 18:33 | |
-!- n4nd0 [~nando@n186-p150.kthopen.kth.se] has quit [Quit: leaving] | 18:43 | |
-!- gsomix [~Miranda@188.168.128.85] has joined #shogun | 19:22 | |
gsomix | konbanwa :) | 19:26 |
@sonney2k | konitschiwa! | 19:33 |
gsomix | sonney2k: so, about one hour before berserk vodka drinking time | 19:52 |
-!- gsomix [~Miranda@188.168.128.85] has quit [Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org] | 20:33 | |
-!- gsomix [~Miranda@188.168.3.84] has joined #shogun | 20:42 | |
-!- aegis1 [aegis1@123.237.52.165] has joined #shogun | 20:48 | |
@sonney2k | gsomix, yay! we are in! | 20:56 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 21:12 | |
n4nd0 | sonney2k: so what kind of news do we have? | 21:13 |
@sonney2k | https://www.google-melange.com/gsoc/accepted_orgs/google/gsoc2013 | 21:13 |
n4nd0 | wow | 21:13 |
n4nd0 | good news indeed | 21:13 |
n4nd0 | yeah the page was loading while I asked :) | 21:13 |
@sonney2k | n4nd0, please twitter about it and blog about it! | 21:13 |
@sonney2k | n4nd0, I am changing the website | 21:14 |
n4nd0 | sonney2k: updating the we are planning to participate to we are participating or sth else? | 21:14 |
@sonney2k | n4nd0, add another message - we are proud to be part of GSoC 2013! | 21:15 |
@sonney2k | n4nd0, but I meant your own blog :D | 21:15 |
n4nd0 | sonney2k: sure, I am on it | 21:15 |
n4nd0 | I am writting in shogun's twitter, then blog, then g+ | 21:15 |
n4nd0 | my browser is only charging 25 accepted orgs and shogun is among them, we must be really important :D | 21:16 |
-!- blackburn [~blackburn@188.168.3.84] has joined #shogun | 21:16 | |
n4nd0 | blackburn: vodking! | 21:16 |
blackburn | uh | 21:16 |
blackburn | just now? | 21:16 |
@sonney2k | https://www.google-melange.com/gsoc/accepted_orgs/google/gsoc2013 | 21:16 |
gsomix | sonney2k: wohooo! | 21:16 |
blackburn | I thought a hour more | 21:17 |
blackburn | :D | 21:17 |
-!- anthrax512 [5e29e2bf@gateway/web/freenode/ip.94.41.226.191] has joined #shogun | 21:19 | |
n4nd0 | twitter doesn't trust I am human... interesting | 21:19 |
@sonney2k | blackburn, http://shogun-toolbox.org/ | 21:19 |
@sonney2k | updated :D | 21:19 |
@sonney2k | blackburn, please write some announcement on the mailinglist... | 21:20 |
blackburn | sonney2k: ok | 21:20 |
anthrax512 | Hi to all, I'm looking for a mentor to GSoC | 21:21 |
@sonney2k | anthrax512, then tell us what you want to do and we will find you someone... | 21:21 |
-!- rishabh [~rishabh@115.249.18.25] has joined #shogun | 21:22 | |
blackburn | haha oh my god that's fast | 21:23 |
-!- jayrambhia [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 21:23 | |
anthrax512 | my main skill is Python, so I can do the "Develop interactive machine learning demos" and "mloss.org and mldata.org integration" tasks | 21:23 |
@sonney2k | blackburn, there is a reason for that - currently only ~20 accepted orgs are shown | 21:23 |
@sonney2k | simply because they haven't all submitted their profile page yet | 21:23 |
@sonney2k | so we appear on the top | 21:23 |
blackburn | sonney2k: >30 already | 21:24 |
blackburn | 42 | 21:24 |
n4nd0 | aah the profile thing | 21:24 |
@sonney2k | anthrax512, then I would suggest you start to familiarize with shogun and read the task description and start doing some simple stuff already | 21:24 |
@sonney2k | blackburn, 20 minutes back it was ~10 | 21:25 |
@sonney2k | blackburn, once you have sent the email have a look at what I've written for shogun | 21:25 |
@sonney2k | blackburn, maybe you have some ideas how to improve... | 21:25 |
blackburn | sonney2k: written wherE? | 21:25 |
@sonney2k | blackburn, didn't you get the acceptance email? | 21:26 |
blackburn | sonney2k: I did | 21:26 |
@sonney2k | the first link | 21:26 |
anthrax512 | @sonney2k, and then I should show some results in the end of april, and you'll make your decision, right? | 21:27 |
blackburn | sonney2k: You need to be a organization administrator for Shogun Machine Learning Toolbox to access this page. | 21:27 |
@sonney2k | blackburn, well log in with your *new* shogun thingy | 21:32 |
@sonney2k | not with your old student account | 21:32 |
blackburn | sonney2k: yeah probably that's the reason | 21:32 |
@sonney2k | blackburn, if that doesn't work I will ask at #gsoc | 21:33 |
blackburn | sonney2k: I think you need to add me there somehow | 21:34 |
blackburn | sonney2k: when I open 'my connections' I see nothing | 21:34 |
@sonney2k | anthrax512, it usually works like this: students check out what they are interested in and already now familiarize with the task and start contributing(!). Then you write a proposal and then it will be ranked by all mentors. | 21:34 |
@sonney2k | blackburn, I also have no connections | 21:34 |
@sonney2k | blackburn, but do you see shogun? | 21:34 |
blackburn | sonney2k: I see only application | 21:34 |
@sonney2k | blackburn, yes then please open the link in the email | 21:35 |
@sonney2k | the first one | 21:35 |
blackburn | The organization with the ID shogun already exists and hence you cannot create a new organization profile for the same ID. If you are actually looking for editing this organization profile please click here. | 21:35 |
blackburn | You need to be a organization administrator for Shogun Machine Learning Toolbox to access this page. | 21:35 |
blackburn | sonney2k: ^ | 21:35 |
@sonney2k | blackburn, what is your link id? | 21:35 |
@sonney2k | and please paste the link you click on | 21:35 |
blackburn | sonney2k: http://www.google-melange.com/gsoc/profile/organization/google/gsoc2013/shogun | 21:35 |
-!- anthrax512 [5e29e2bf@gateway/web/freenode/ip.94.41.226.191] has quit [Quit: Page closed] | 21:35 | |
-!- anthrax512 [5e29e2bf@gateway/web/freenode/ip.94.41.226.191] has joined #shogun | 21:36 | |
@sonney2k | blackburn, waht is your link id again? | 21:42 |
blackburn | sonney2k: username? | 21:42 |
@sonney2k | blackburn, link id | 21:42 |
@sonney2k | yes username | 21:42 |
@sonney2k | (maybe) | 21:42 |
blackburn | sonney2k: sergey_lisitsyn | 21:42 |
@sonney2k | blackburn, I asked carols | 21:44 |
@sonney2k | btw 177 organizations in | 21:44 |
@sonney2k | blackburn, carols said I should ask in melange | 21:45 |
@sonney2k | #melange | 21:45 |
@sonney2k | hmmhh | 21:46 |
@sonney2k | blackburn, did you use that username in last years too? | 21:46 |
blackburn | sonney2k: yes I didn't change anything | 21:46 |
@sonney2k | too bad | 21:48 |
@sonney2k | then that certainly is the problem | 21:48 |
@sonney2k | hmmhh I am getting email from students already... | 21:49 |
blackburn | sonney2k: oh you will get a lot | 21:49 |
blackburn | due to simplicity of some tasks | 21:49 |
@sonney2k | random forests... | 21:50 |
blackburn | sonney2k: ahh.. | 21:50 |
@sonney2k | looks like it will take the melange guys some time | 21:50 |
blackburn | sonney2k: is that a problem? we can wait a little, right? | 21:51 |
@sonney2k | blackburn, well announce the thing nevertheless - you can do edits later if you are not happy | 21:54 |
@sonney2k | blackburn, that is what I wrote | 21:54 |
@sonney2k | https://www.google-melange.com/gsoc/org/google/gsoc2013/shogun | 21:54 |
blackburn | sonney2k: I am quite ok with it | 21:54 |
@sonney2k | blackburn, n4nd0 I also updated http://shogun-toolbox.org/page/Events/workshop2013 | 21:54 |
@sonney2k | blackburn, btw chris jsut told me he won't be at the workshop :/ | 21:54 |
n4nd0 | good | 21:54 |
blackburn | sonney2k: did you receive anything from the big boss? | 21:55 |
n4nd0 | oh | 21:55 |
n4nd0 | I meant good for the update, not for Chris news :S | 21:55 |
blackburn | sonney2k: oh bad I won't met him | 21:55 |
n4nd0 | sonney2k: what about changing planning to partipate to partipating http://shogun-toolbox.org/page/Events/gsoc2013 | 21:55 |
blackburn | n4nd0: sure please change ;) | 21:55 |
n4nd0 | blackburn: let me fix my computer first please :S | 21:56 |
n4nd0 | I am even scared | 21:56 |
n4nd0 | idk what the hell I have done | 21:56 |
blackburn | n4nd0: ahh yeah sure | 21:56 |
@sonney2k | n4nd0, no add a new news item | 21:56 |
@sonney2k | saying that we are in | 21:56 |
n4nd0 | sonney2k: ok | 21:56 |
@sonney2k | keep the other :D | 21:56 |
@sonney2k | n4nd0, like here http://shogun-toolbox.org/page/Events/gsoc2012 | 21:56 |
@sonney2k | even with the image :D | 21:56 |
n4nd0 | sure hehe | 21:57 |
-!- zxtx [~zv@76.91.177.87] has joined #shogun | 21:57 | |
-!- anthrax512 [5e29e2bf@gateway/web/freenode/ip.94.41.226.191] has quit [Quit: Page closed] | 21:59 | |
n4nd0 | three hurrays for me, I created a new group in my system for printer rights | 21:59 |
n4nd0 | and add myself to the group with usermode -G blah blah | 21:59 |
n4nd0 | usermod* | 21:59 |
-!- katia [d5578f7b@gateway/web/freenode/ip.213.87.143.123] has joined #shogun | 22:01 | |
-!- ahcorde [553324b1@gateway/web/freenode/ip.85.51.36.177] has joined #shogun | 22:06 | |
-!- katia [d5578f7b@gateway/web/freenode/ip.213.87.143.123] has quit [Quit: Page closed] | 22:06 | |
@sonney2k | n4nd0, and? | 22:07 |
-!- aegis1 [aegis1@123.237.52.165] has quit [] | 22:07 | |
n4nd0 | sonney2k: it seems there is an option called append I should have used | 22:08 |
n4nd0 | sonney2k: now I am happily just belonging to the group for printers | 22:08 |
n4nd0 | I cannot sudo anything.... | 22:08 |
@sonney2k | n4nd0, why didn't you edit /etc/groups manually? | 22:08 |
n4nd0 | I thought it was probably better to do it using the command | 22:09 |
n4nd0 | if any other file needed to be updated | 22:09 |
n4nd0 | I am young and stupid :'( | 22:10 |
-!- katia [d5578f7b@gateway/web/freenode/ip.213.87.143.123] has joined #shogun | 22:18 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds] | 22:18 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:20 | |
@sonney2k | blackburn, well done | 22:23 |
@sonney2k | so we have planet d.o / g+ / twitter / mailinglist announcemnts | 22:24 |
@sonney2k | blackburn, maybe we should send an email to machine learning news too? | 22:25 |
@sonney2k | in any case what is missing is the workshop call for papers | 22:25 |
blackburn | sonney2k: debian have the same thing with backup | 22:27 |
blackburn | sonney2k: see #melange | 22:27 |
-!- bogdanc [~bogdan@p16.eregie.pub.ro] has joined #shogun | 22:27 | |
-!- praveen97uma [ab4ce0b0@gateway/web/freenode/ip.171.76.224.176] has joined #shogun | 22:28 | |
@sonney2k | blackburn, try again | 22:30 |
@sonney2k | I've re-added you | 22:30 |
blackburn | sonney2k: received email | 22:30 |
@sonney2k | or tried at least | 22:30 |
@sonney2k | check if you can access the link | 22:30 |
@sonney2k | blackburn, btw please email each and every mentor telling them how to register... | 22:31 |
-!- jayrambhia [73f88294@gateway/web/freenode/ip.115.248.130.148] has left #shogun [] | 22:31 | |
blackburn | ohh yeah ok | 22:31 |
@sonney2k | blackburn, can you finally access the page? | 22:32 |
bogdanc | hi! I would be interested in participating in GSOC working on a shogun project, but I do not have any ML knowleadge now. Though, I am willing to learn about it. Do you think that is it possible for me to solve one task in this condition? | 22:33 |
-!- rishabh [~rishabh@115.249.18.25] has quit [Ping timeout: 252 seconds] | 22:33 | |
n4nd0 | bogdanc: Yeah, sure. There are a couple of tasks that do not require much specific ML learning. | 22:33 |
blackburn | sonney2k: still no | 22:34 |
n4nd0 | ML knowledge more than learning hehe | 22:34 |
-!- rishabh [~rishabh@202.78.172.162] has joined #shogun | 22:34 | |
naywhayare | hello there. long-time lurker and maintainer of another ML library that got accepted into GSoC here. quick question: how many applications am I going to end up with? 1/10/100/1000/10 billion? | 22:34 |
blackburn | naywhayare: hah congrats! | 22:34 |
naywhayare | I've not done GSoC before so I have no idea what to expect. any help is appreciated :) | 22:34 |
blackburn | naywhayare: yeah about 10-12 billion | 22:34 |
blackburn | ;) | 22:34 |
naywhayare | really? I wouldn't be too surprised. I'm trying to figure out what level of filtering I'll need to be doing | 22:35 |
blackburn | naywhayare: we received a bit more than 100 last year I think | 22:35 |
bogdanc | so I have no real chance with a task involving ML? | 22:35 |
@sonney2k | blackburn, IIRC 50-60 | 22:35 |
blackburn | sonney2k: oh my mistake | 22:36 |
naywhayare | oh wow, okay. our humble little outfit surely can't handle any more than 10 (I suspect my other "mentors" are in it for the free t-shirts) | 22:36 |
blackburn | naywhayare: what is your org? | 22:36 |
blackburn | bogdanc: there is always a chance | 22:37 |
naywhayare | blackburn: mlpack | 22:37 |
blackburn | naywhayare: nice website! | 22:38 |
blackburn | :) | 22:38 |
naywhayare | I need to polish up the application page quite a lot (judging by the other pages I've looked at) but I got something up there "just for now" | 22:38 |
naywhayare | thanks. I wanted to keep it super-minimal, but in some respects it's *too* minimal | 22:38 |
bogdanc | I know that there is a chance, but I don't want to set targets I can't reach. that's why I asked you who have much more experience in the field | 22:38 |
@sonney2k | naywhayare, it's like windows 8 isn't it :D | 22:39 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:39 | |
@sonney2k | naywhayare, since you are in too - maybe we can have some synergies | 22:39 |
naywhayare | haha. "looks great on mobile devices" | 22:39 |
blackburn | bogdanc: well ML is not that hard | 22:39 |
@sonney2k | naywhayare, at least with say being able to read your / our data formats or exchanging simple models | 22:40 |
naywhayare | sonney2k: if you have any ideas, I'd be glad to work together. in some aspects shogun and mlpack are orthogonal | 22:40 |
n4nd0 | bogdanc: Sure. Take a look to the ideas page and see what would be suitable for you. For example I think you don't need to be a ML guru for the demos project http://www.shogun-toolbox.org/page/Events/gsoc2013_ideas#demos | 22:40 |
naywhayare | ah, now there is a good idea. we haven't settled on a saved model format completely yet (rudimentary efforts are xml) | 22:40 |
n4nd0 | bogdanc: and it may be a good fit to starting learning about it. | 22:40 |
naywhayare | most of the tasks mlpack does are one-time-use models so there's not much use saving a model | 22:41 |
blackburn | naywhayare: I also have a template C++ library on dimension reduction if you are interested :) | 22:41 |
naywhayare | can you point me towards it? I'd be interested in poking at it | 22:42 |
blackburn | naywhayare: https://github.com/lisitsyn/tapkee | 22:42 |
naywhayare | just wish I had more time to really look into all the things I wanted to :-S | 22:42 |
blackburn | naywhayare: it is the core of dimension reduction in shogun | 22:43 |
deerishi | any suggestions on projects involving ICA/BSS and signal procesing? | 22:43 |
blackburn | deerishi: suggestions like? | 22:43 |
naywhayare | ah, it has the ridiculously ugly john langford cover tree code | 22:43 |
blackburn | naywhayare: it works.. :) | 22:43 |
naywhayare | it does, but it ain't pretty :) | 22:43 |
n4nd0 | deerishi: There is actually a project idea on ICA and BSS :) Not sure if you meant that. | 22:43 |
deerishi | i meant that | 22:44 |
deerishi | :) | 22:44 |
n4nd0 | deerishi: aham! All right. | 22:44 |
blackburn | naywhayare: I have had a bad experience with 'prettier' covertree https://github.com/DNCrane/Cover-Tree | 22:44 |
@sonney2k | pretty != working != fast | 22:45 |
blackburn | sonney2k: my other code is much prettier and working and fast ;) | 22:45 |
n4nd0 | deerishi: mmm so regarding ICA, I know there is some open source implementations of ICA strategies (in sklearn for example). You could check them out and try to get something into Shogun as an initial patch. | 22:45 |
naywhayare | my thought has been that a lot of the stuff jl did, while perhaps faster, is only incrementally (i.e. 1-3%) faster than a much prettier counterpart | 22:45 |
naywhayare | the fact that it's not documented at all is also a big problem. it took me forever to realize what he was doing with the empty lists to avoid mallocs | 22:46 |
@sonney2k | blackburn, that is also interesting https://google-melange.appspot.com/gsoc/org/google/gsoc2013/pagmo | 22:46 |
n4nd0 | blackburn, naywhayare, sonney2k : why do people dislike JL implementations? :))) | 22:46 |
@sonney2k | n4nd0, I don't | 22:46 |
blackburn | n4nd0: because they are .. let me cite naywhayare .. 'ridiculiously ugly' | 22:46 |
n4nd0 | hehe | 22:46 |
naywhayare | n4nd0: 'cause they don't implement what's in the paper, they're ugly, undocumented, unmaintained, and not particularly usable in other settings | 22:46 |
naywhayare | also not generalized at all and specific to only one task | 22:47 |
naywhayare | ...I have *particularly* strong opinions about jl code | 22:47 |
-!- medeeiip [~medeeiip@103.2.132.232] has joined #shogun | 22:47 | |
n4nd0 | naywhayare: hehe I agree with you in that it is rather hard to go through | 22:47 |
naywhayare | the funniest part is that in my experience, I have yet to find a dataset where cover trees perform faster than well-written kd trees (the construction cost is usually huge) | 22:47 |
blackburn | I have to second that guys | 22:47 |
@sonney2k | naywhayare, heh - I guess the same I have about TJs code (joachims...) | 22:47 |
naywhayare | curiously the cover tree paper did not include comparisons with kd-trees. | 22:47 |
bogdanc | I have noticed that one too as being an easier one. but how about something under machine learning tasks? | 22:47 |
-!- praveen97uma_ [ab4ce0b0@gateway/web/freenode/ip.171.76.224.176] has joined #shogun | 22:47 | |
blackburn | naywhayare: oh there is a strong reason why I use covertree | 22:48 |
blackburn | it is a metric tree | 22:48 |
n4nd0 | bogdanc: mmm sure, check them out and see what you like | 22:48 |
@sonney2k | bogdanc, and maybe try out some graphical / interactive examples of shogun to see whats going on | 22:49 |
-!- praveen97uma [ab4ce0b0@gateway/web/freenode/ip.171.76.224.176] has quit [Ping timeout: 245 seconds] | 22:49 | |
blackburn | naywhayare: as I needed something that works with arbitrary distance/kernel it was very important for me | 22:49 |
naywhayare | this is true | 22:49 |
naywhayare | although in that case I do suspect ball trees would still be faster than cover trees | 22:49 |
naywhayare | but I haven't thoroughly taken a look | 22:49 |
blackburn | naywhayare: yeah I just didn't get proper ball tree code yet | 22:50 |
blackburn | :) | 22:50 |
bogdanc | so with enough time and energy invested, it is faisable for a beginner to solve one of those, right? | 22:50 |
naywhayare | :) | 22:50 |
-!- deerishi_ [73f88294@gateway/web/freenode/ip.115.248.130.148] has joined #shogun | 22:50 | |
n4nd0 | bogdanc: Yeah, I think so. | 22:50 |
-!- deerishi [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Ping timeout: 245 seconds] | 22:50 | |
-!- katia [d5578f7b@gateway/web/freenode/ip.213.87.143.123] has quit [Quit: Page closed] | 22:50 | |
n4nd0 | be careful with the one called "Implement estimators of large-scale sparse Gaussian densities" | 22:51 |
deerishi_ | n4nd0: thank you will check that out | 22:51 |
blackburn | hahaha | 22:51 |
blackburn | n4nd0: good | 22:51 |
n4nd0 | we were not sure if someone around here will be a suitable student :D | 22:51 |
n4nd0 | including mentors hehe | 22:51 |
@sonney2k | bogdanc, but it certainly depends on your skills (math, coding, ...) | 22:52 |
@sonney2k | blackburn, hmmhh I don't see orange yet | 22:54 |
-!- deerishi_ [73f88294@gateway/web/freenode/ip.115.248.130.148] has quit [Client Quit] | 22:54 | |
@sonney2k | in the accepted list... | 22:54 |
blackburn | how many are in already? | 22:54 |
@sonney2k | blackburn, ahh btw please CC me when you email mentors so that I know - thanks | 22:54 |
@sonney2k | blackburn, ~90 | 22:54 |
bogdanc | ok, thank you very much! | 22:54 |
blackburn | sonney2k: yes sure | 22:54 |
n4nd0 | sonney2k, blackburn : what do you think are the project priorities this year? | 22:55 |
-!- ahcorde [553324b1@gateway/web/freenode/ip.85.51.36.177] has quit [Ping timeout: 245 seconds] | 22:55 | |
@sonney2k | n4nd0, gsoc wise you mean? | 22:55 |
n4nd0 | sonney2k: yes | 22:55 |
blackburn | n4nd0: world domination as usual | 22:55 |
@sonney2k | in general it IMHO it really is accessibility | 22:55 |
@sonney2k | and that :D | 22:55 |
n4nd0 | hehe | 22:55 |
medeeiip | hello everyone I'm new to this project and want to build shogun on windows. I can see a cmakeList in /src but it is creating a sln file with a project only with the cpp files. How can I get also the .h files inside the project? | 22:56 |
n4nd0 | mmm I think we have an errata in the ideas list | 22:56 |
@sonney2k | n4nd0, now that we have proper infrastructure we should be able to generate some nice applications / doc :D | 22:56 |
@sonney2k | n4nd0, not so unlikely... | 22:57 |
@sonney2k | medeeiip, shogun & windows is tough - cygwin is the only thing I ever got to work | 22:57 |
n4nd0 | sonney2k: the second project in accesibility list. We removed from index but not from descriptions | 22:57 |
n4nd0 | medeeiip: Unfortunately cmake integration in shogun is far from complete. I believe someone started sth, but nothing else. | 22:58 |
@sonney2k | n4nd0, indeed this format reading stuff is covered in fas reading and writing shogun features | 22:58 |
blackburn | naywhayare: so let me know if you would like to integrate my stuff - I will assist all the things if so | 22:59 |
medeeiip | If I create a new cmakelist is it possible ? I mean to say is there any more additional dependency? | 22:59 |
@sonney2k | medeeiip, if you get it to work we would be most happy to use it! | 22:59 |
@sonney2k | medeeiip, and merge your patch | 22:59 |
n4nd0 | sonney2k: let's take it out then. Are you editing or should I? | 22:59 |
@sonney2k | n4nd0, please do | 22:59 |
@sonney2k | medeeiip, well there are a couple but there are no hard dependencies in shogun | 23:00 |
n4nd0 | blackburn: you are not changing anything in ideas either,are you? | 23:00 |
blackburn | n4nd0: no I do not change anything | 23:00 |
naywhayare | blackburn: I'll take a look, but I have a ton on my plate. I'll probably have a chance to look into it in a couple weeks | 23:01 |
blackburn | naywhayare: yes sure - just let me know if you (would like) and we will discuss it then | 23:03 |
-!- rishabh [~rishabh@202.78.172.162] has quit [Ping timeout: 264 seconds] | 23:03 | |
naywhayare | blackburn: one interesting thing to think about for DR may be this: | 23:04 |
naywhayare | some DR methods like MVU are based on semidefinite programs (SDPs) | 23:04 |
naywhayare | about ten years ago there was some work by Monteiro and Burer which reformulated SDPs as low-rank SDPs | 23:05 |
blackburn | naywhayare: yeah may be I'll allocate a gsoc task for that as I lack MVU | 23:05 |
naywhayare | which are not convex like SDPs, so you don't get to use all the nice convex optimization theory, but in practice they run faster (i.e. O(N^2 log N) not O(N^3)) and give solutions nearly as good | 23:05 |
naywhayare | so applying LRSDPs to MVU has reportedly been done by my "predecessor" (Vasiloglou... the paper is out there somewhere) | 23:06 |
naywhayare | but I haven't ever managed to reproduce the results though I have tried quite extensively | 23:06 |
blackburn | good to know, thanks naywhayare | 23:06 |
naywhayare | a working LRSDP implementation applied to Weinberger's usual 'formulate-it-as-an-SDP' works (and other people who do the same) could provide some really good speedup for that | 23:06 |
naywhayare | I have a working LRSDP implementation in mlpack (at least, it passes the simple tests) but I could never get it to work for any ML task | 23:07 |
naywhayare | anyway, perhaps approaching it again from a fresh start would yield better results | 23:07 |
-!- praveen97uma_ [ab4ce0b0@gateway/web/freenode/ip.171.76.224.176] has quit [Ping timeout: 245 seconds] | 23:19 | |
gsomix | time to read ideas page | 23:34 |
-!- phd [lvoronova1@77-120-201-195.sumy.volia.net] has joined #shogun | 23:38 | |
-!- praveen97uma [6adb6aa5@gateway/web/freenode/ip.106.219.106.165] has joined #shogun | 23:44 | |
-!- nando [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 23:46 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 23:46 | |
-!- nando is now known as n4nd0 | 23:46 | |
n4nd0 | yuhu, I fixed it now :) | 23:47 |
-!- phd [lvoronova1@77-120-201-195.sumy.volia.net] has quit [] | 23:47 | |
-!- phoenix_ukr [lvoronova1@77-120-201-195.sumy.volia.net] has joined #shogun | 23:48 | |
-!- praveen97uma [6adb6aa5@gateway/web/freenode/ip.106.219.106.165] has quit [Ping timeout: 245 seconds] | 23:53 | |
-!- phoenix_ukr [lvoronova1@77-120-201-195.sumy.volia.net] has quit [] | 23:55 | |
-!- phoenix_ukr [lvoronova1@77-120-201-195.sumy.volia.net] has joined #shogun | 23:55 | |
-!- gggggggggggggg [lvoronova1@77-120-201-195.sumy.volia.net] has joined #shogun | 23:56 | |
-!- gggggggggggggg [lvoronova1@77-120-201-195.sumy.volia.net] has quit [Client Quit] | 23:56 | |
-!- praveen97uma [6adb6aa5@gateway/web/freenode/ip.106.219.106.165] has joined #shogun | 23:56 | |
--- Log closed Tue Apr 09 00:00:10 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!