IRC logs of #shogun for Wednesday, 2012-05-23

--- Log opened Wed May 23 00:00:41 2012
-!- gsomix [~gsomix@188.168.14.91] has quit [Quit: Ex-Chat]00:08
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]00:24
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun00:28
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host]00:28
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun00:28
-!- blackburn [~blackburn@188.122.250.167] has quit [Quit: Leaving.]00:54
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has left #shogun []01:07
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]01:51
-!- av3ngr [av3ngr@nat/redhat/x-umdlpjzkblrvncrc] has joined #shogun02:24
-!- av3ngr [av3ngr@nat/redhat/x-umdlpjzkblrvncrc] has quit [Read error: Connection reset by peer]02:45
-!- av3ngr [av3ngr@nat/redhat/x-tebxvkyvuqhuydkb] has joined #shogun02:46
-!- abn_ [av3ngr@nat/redhat/x-fsheujvuukzufsvw] has joined #shogun03:41
-!- av3ngr [av3ngr@nat/redhat/x-tebxvkyvuqhuydkb] has quit [Ping timeout: 265 seconds]03:45
-!- av3ngr [av3ngr@nat/redhat/x-ngxvzgqzujorajln] has joined #shogun03:51
-!- abn_ [av3ngr@nat/redhat/x-fsheujvuukzufsvw] has quit [Ping timeout: 252 seconds]03:55
-!- av3ngr [av3ngr@nat/redhat/x-ngxvzgqzujorajln] has quit [Read error: Connection reset by peer]04:33
-!- av3ngr [av3ngr@nat/redhat/x-hetklmjjpgljcpum] has joined #shogun04:33
-!- av3ngr [av3ngr@nat/redhat/x-hetklmjjpgljcpum] has quit [Read error: Connection reset by peer]04:44
-!- av3ngr [av3ngr@nat/redhat/x-gyynhgqyqjwjffil] has joined #shogun04:45
-!- pluskid [~pluskid@li400-235.members.linode.com] has joined #shogun04:56
-!- pluskid [~pluskid@li400-235.members.linode.com] has quit [Client Quit]04:56
-!- pluskid [~pluskid@111.120.11.156] has joined #shogun05:01
-!- abn_ [av3ngr@nat/redhat/x-ixsihisabqgorgry] has joined #shogun05:12
-!- abn_ [av3ngr@nat/redhat/x-ixsihisabqgorgry] has quit [Client Quit]05:13
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]05:13
-!- pluskid [~pluskid@111.120.11.156] has quit [Ping timeout: 252 seconds]05:17
-!- av3ngr [av3ngr@nat/redhat/x-gyynhgqyqjwjffil] has quit [Ping timeout: 276 seconds]05:17
-!- pluskid [~pluskid@li400-235.members.linode.com] has joined #shogun05:17
-!- vikram360 [~vikram360@117.192.161.189] has joined #shogun07:01
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun07:15
@sonney2kpluskid, wiking did you manage to git rebase or to this filter-branch'd repository?08:09
CIA-113shogun: Chiyuan Zhang master * r2c0ea33 / examples/undocumented/python_modular/classifier_multiclass_ecoc.py : exclude abstract class from testing. - http://git.io/wV23Kw08:14
CIA-113shogun: Chiyuan Zhang master * r1983da0 / examples/undocumented/python_modular/evaluation_clustering.py : Fix evaluation_clustering error. - http://git.io/FCthiQ08:14
CIA-113shogun: Chiyuan Zhang master * rd3e4418 / .gitignore : added temp files generated when running python_modular examples to gitignore - http://git.io/FhWKbw08:14
CIA-113shogun: Soeren Sonnenburg master * r3f61978 / (3 files in 2 dirs): Merge pull request #549 from pluskid/multiclass - http://git.io/tHxyhQ08:14
pluskidsonney2k: what filter-branch'd repo? I did a completely new fork as stated in the mail08:18
@sonney2kpluskid, yes exactly that is what you had to do08:18
@sonney2kbut your pull request needs to be 'rebased' I guess format-patch is the only option08:19
pluskidsonney2k: which pull request? You just merged them, as CIA-113 said above08:19
@sonney2kpluskid, I merged one08:20
@sonney2knot the other08:20
pluskidthe other ? let me see08:20
pluskidoh... just forget about that one08:21
pluskidI guess I have re-write the code in that pull request08:22
pluskidI deleted my previous fork repo and thought I lost them08:22
@sonney2kpluskid, ok08:28
@sonney2kpluskid, btw do you know these 2 cool multiclass papers?08:28
pluskidwhich ones?08:29
pluskidI'm doing a survey these days08:29
@sonney2k"ShareBoost: Efficient Multiclass Learning with08:30
@sonney2kFeature Sharing" by Shalev-Shwartz et al"ShareBoost: Efficient Multiclass Learning with08:30
@sonney2kFeature Sharing" by Shalev-Shwartz et al08:30
pluskidoh, this one is mentioned in the GSoC idea page of shogun :D08:31
@sonney2k"Fast and Balanced: Efficient Label Tree Learning for08:31
@sonney2kLarge Scale Object Recognition" by Deng et al08:31
pluskidthis one, too08:31
pluskidthanks all the same!08:32
@sonney2kpluskid, both are nips papers from last years NIPS08:32
@sonney2kyes these two08:32
@sonney2kI talked to the authors last year and I think these could be really useful08:32
pluskidyeah, I did brief scanning of them before, I'll try to dig deeper this week08:33
pluskidsonney2k: oh, cool!08:33
@sonney2kpluskid, I don't know how far you are with the general ECOC library conversion08:33
@sonney2kmaybe these 2 are good candidates for july or so08:34
pluskidsonney2k: I have translated most part08:34
pluskidof ecoclib08:34
@sonney2kthen june :D08:34
@sonney2kpluskid, learning ECOC is not really helping08:34
@sonney2kgunnar (from shoGUN) tried that once08:34
@sonney2kbut it didn't help08:34
@sonney2kor almost not08:34
pluskidhmm08:34
@sonney2knips paper still...08:34
pluskidI have the same feeling (though I didn't do really rigious experiments)08:35
@sonney2kpluskid, so *I* would rather go for other promising MC approaches08:35
@sonney2kalright08:35
@sonney2kgot to leave train08:35
@sonney2kbrb08:35
pluskidyeah, I'm doing a survey to collect and category related algorithms08:35
pluskidI'll also discuss with Cheng the next week08:35
pluskidafter I finished a rough survey08:35
pluskidabout which algorithm to go next08:36
-!- crazy_horse [~sonnenbu@194.78.35.195] has joined #shogun08:50
crazy_horseok I am back08:51
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun08:51
-!- uricamic [~uricamic@2001:718:2:1634:fda6:1b8d:2f96:ced1] has joined #shogun09:08
-!- crazy_horse is now known as sonne|work09:22
CIA-113shogun: Soeren Sonnenburg master * re59e42f / (.gitignore src/interfaces/modular/Machine.i): %rename apply of any type and include Makefile in gitignore - http://git.io/R5dMMA09:41
sonne|workpluskid: the other idea I had is to look at the LSHTC09:46
sonne|workpluskid: this one http://lshtc.iit.demokritos.gr/09:46
sonne|workthey had lots of submissions / papers last / this year09:47
pluskidsonne|work: cool09:47
sonne|workbut they usually heavily make use of structure in the data (hierachies aka trees)09:47
pluskidhmm, so we'd better also remind n4nd009:48
sonne|workso maybe the winning entry from last year might be interesteing09:48
sonne|workpluskid: but it is not SO - right?09:48
sonne|workmulticlass/multitask09:48
pluskidoh, you mean the structure in X?09:48
sonne|workpluskid: no in y09:48
sonne|workyou know that there is a hierarchy09:49
pluskidthen that belongs to SO I guess09:49
sonne|workI guess you can see it that way - or you can see it the ECOC or so way09:49
sonne|workhaving trees of classifiersx09:49
sonne|workmatching the hierarchy09:50
pluskidhmm, that makes sense09:50
-!- nickon [~noneedtok@kotnet-147.kulnet.kuleuven.be] has joined #shogun09:50
pluskidmulticlass/multitask/multilabel...09:51
n4nd0sonne|work: oh, how did that happen to the pull request? I don't remember I pushed any changes yesterday10:19
n4nd0pluskid: hey!10:21
n4nd0about this issue of doing the fork again, did you manage to rebase or have you just forked the repo again?10:22
pluskidn4nd0: I forked again10:23
pluskidthings a totally messy10:23
pluskidI don't know what they have done yesterday10:23
pluskidthey are horrible crazy horses :D10:24
n4nd0:D10:25
sonne|workthe crazy_{horses,russians} recreated the whole repository10:26
sonne|workeverything is invalid basically10:26
n4nd0O_O that sounds pretty bad10:27
sonne|workwe had to do this because of blackburns accidental merge some huuuuge data sets10:27
n4nd0I see10:27
sonne|workn4nd0: just do a fresh clone10:27
sonne|workand then everything is good again10:28
sonne|workshogun .git dir is now ~30MB10:28
n4nd0yeah, I am on my way :)10:28
sonne|worknot 158 :)10:28
n4nd0that's nice10:28
n4nd0but I need to delete my fork in github as well, right?10:39
pluskidI did that10:40
n4nd0there's no problem with unmerged branches?10:40
pluskidI think they will lost if you delete the repo10:40
n4nd0I have my branch so that has not been merged into the repo10:40
n4nd0yeah ... that's too bad :S10:40
pluskidmaybe you have to backup them with git-format-patch and re-apply them when the new fork is created10:41
n4nd0I was thinking of that10:41
n4nd0kind of *not very* comfortable though10:41
n4nd0but anyway ...10:41
pluskiddon't know of any easier way10:41
n4nd0sonne|work: do you know any alternative?10:42
n4nd0otherwise, I will start doing it10:42
sonne|workn4nd0: well that is pretty easy imho10:43
sonne|workuse git format patch10:43
sonne|workyou don't need to delete your fork you need to force push at some point10:43
sonne|workbut keep a copy of what you had before...10:43
n4nd0sure10:44
pluskidany in the university network with springer subscription help me to download this paper?10:45
pluskidhttp://www.springerlink.com/content/n30xeen83d81ylbu/?MUD=MP10:45
pluskidThanks!10:45
sonne|workpluskid: http://www.google.be/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CEgQFjAA&url=http%3A%2F%2Fhunch.net%2F~beygel%2Fsecoc.ps&ei=gKS8T4OGDtSm8gPbl5gj&usg=AFQjCNHBUHtQ01KMZ3320-S9YJSLLvCwyw10:49
sonne|workpluskid: or just google the title10:49
sonne|workyou will find it at hunch10:49
sonne|work(JL's homepage)10:49
pluskidsonne|work: ah, thanks10:50
pluskidI had this .ps file, but fonts looks a bit strange10:50
pluskidin my box10:50
pluskidso I'm seeking a PDF10:51
sonne|workpluskid: there is also the .pdf10:51
sonne|workIIRC second google hit10:51
pluskidOK, I'll google it10:51
sonne|workpluskid: yes this one http://www.google.be/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=0CFcQFjAB&url=http%3A%2F%2Fhunch.net%2F~jl%2Fprojects%2Freductions%2Fpecoc%2Ffinal%2Fsecoc.pdf&ei=FKW8T8iNEpTx8QOz38Fl&usg=AFQjCNHITbB2AYcizO_r31aO00FB8KcLzA10:51
pluskidhaha, that is the one with strange font10:52
pluskidtoo10:52
pluskidanyway, I will go with this10:53
pluskidbut if anyone have access to the COLT official copy, it would be nicer10:53
n4nd0pluskid: I have tried over ssh to my university but I got ERROR 403: Forbidden, sorry10:54
pluskidn4nd0: OK, thanks anyway!10:54
n4nd0I can try later when I get there in any case again10:54
n4nd0I have never found a problem to get a paper being there10:54
pluskidn4nd0: cool!10:54
n4nd0John Langford is the author10:57
n4nd0it seems that he's done lot of stuff!!10:57
pluskidyes!10:59
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]10:59
pluskidJL has lots of MC stuff, some recent work involves tree style prediction for huge scale data11:02
-!- pluskid [~pluskid@li400-235.members.linode.com] has quit [Ping timeout: 276 seconds]11:30
-!- pluskid [~pluskid@111.120.11.156] has joined #shogun11:46
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun11:56
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun12:03
-!- karlnapf [~heiko@host86-171-220-173.range86-171.btcentralplus.com] has joined #shogun12:26
-!- vikram360 [~vikram360@117.192.161.189] has quit [Ping timeout: 265 seconds]12:31
n4nd0by the way, what about the warnings in DynamicArray and SGMatrix12:33
n4nd0should we get rid of them?12:33
blackburnwhich warnings?12:34
pluskidoperator[] ?12:34
blackburnah it should be ignored12:35
n4nd0why?12:35
n4nd0I get *a lot* of ../../shogun/lib/DynamicArray.h:296: Warning 516: Overloaded method shogun::CDynamicArray< uint64_t >::element(int32_t) ignored,12:35
blackburnn4nd0: ??12:36
n4nd0and also the can't wrap operator ones12:36
blackburnis your git up to date?12:36
n4nd0blackburn: yes12:36
n4nd0it might be something mine if you don't see them12:36
blackburnn4nd0: iirc gsomix fixed those ones12:39
pluskidsth like this: ../../shogun/lib/SGVector.h:193: Warning 389: operator[] ignored (consider using %extend)12:40
pluskidnot *too* many though12:40
-!- wiking [~wiking@huwico/staff/wiking] has quit [Read error: Operation timed out]12:40
blackburnoperator[] should be ignored in modular I think12:41
blackburnhmm it is already ignored12:50
n4nd0this is part of how it looks for me12:51
n4nd0http://snipt.org/ujfR212:51
n4nd0lot of them are the same errors repeated12:52
n4nd0is that something mine?12:52
n4nd0it is everything new12:52
n4nd0github fork, the local clone, and in master12:52
blackburnthat's strange12:54
n4nd0I am going for lunch, I'll be back later12:59
blackburnok I removed a few13:01
CIA-113shogun: Sergey Lisitsyn master * rba81d43 / (2 files): Ignored {=,*,(),[]}operators in modular - http://git.io/TZ5WGQ13:04
blackburnsonne|work: should I push to shogun-toolbox and google code now too?13:05
-!- wiking [~wiking@2002:4e17:bd70:1:a9ae:ac1:93a7:13a1] has joined #shogun13:14
-!- wiking [~wiking@2002:4e17:bd70:1:a9ae:ac1:93a7:13a1] has quit [Changing host]13:14
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun13:14
-!- sonne|work [~sonnenbu@194.78.35.195] has left #shogun []13:16
n4nd0blackburn: do we have a goodle code?13:26
blackburnyes13:27
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun13:28
-!- karlnapf [~heiko@host86-171-220-173.range86-171.btcentralplus.com] has quit [Ping timeout: 245 seconds]13:37
-!- vikram360 [~vikram360@117.192.161.189] has joined #shogun13:43
-!- blackburn [~blackburn@188.122.250.167] has quit [Ping timeout: 245 seconds]13:47
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]14:27
-!- karlnapf [~heiko@host86-181-81-215.range86-181.btcentralplus.com] has joined #shogun14:35
-!- wiking [~wiking@b255h244.ugent.be] has joined #shogun14:38
-!- wiking [~wiking@b255h244.ugent.be] has quit [Changing host]14:38
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun14:38
-!- wiking [~wiking@huwico/staff/wiking] has quit [Remote host closed the connection]14:38
-!- wiking [~wiking@we02c096.ugent.be] has joined #shogun14:46
-!- wiking [~wiking@we02c096.ugent.be] has quit [Changing host]14:46
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun14:46
-!- vikram360 [~vikram360@117.192.161.189] has quit [Ping timeout: 256 seconds]15:06
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]15:22
-!- vikram360 [~vikram360@117.192.161.189] has joined #shogun15:32
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun15:40
-!- pluskid [~pluskid@111.120.11.156] has quit [Ping timeout: 276 seconds]16:20
-!- pluskid [~pluskid@li400-235.members.linode.com] has joined #shogun16:21
@sonney2kblackburn of course16:33
@sonney2kn4nd0, yes we have 2 mirrors next to github16:33
CIA-113shogun: Heiko Strathmann master * r9725b6d / (10 files): Introduced CQuadraticTimeMMD and another level in the class hierarchy: - http://git.io/_IwAfQ16:34
CIA-113shogun: Heiko Strathmann master * r4b848b4 / (10 files): Merge pull request #550 from karlnapf/master - http://git.io/VZbA5A16:34
@sonney2kyou never know... :)16:34
@sonney2kkarlnapf, yeah please praise me more ... took me hours ...16:34
karlnapfall hail sonney2k for cleaning the git logs :) :) :)16:34
karlnapfyear, shogun is so much faster than matlab16:35
@sonney2kkarlnapf, doing what?16:35
karlnapfcomparing performance of code in matlab and my shogun implementation16:35
karlnapfwell its no surprise, nested loops16:36
@sonney2kkarlnapf, well that is kind of unfair16:38
@sonney2kyou need to compare it to vectorized code :)16:38
karlnapfI know :)16:38
karlnapfsonney2k, if you are interested, have a look into the test classes I made16:39
karlnapfI think the structure is quite nice like this16:39
karlnapfWill add some examples how to use them (ultra easy) soon16:39
karlnapfgotta go now, studying ...16:40
karlnapfcu16:40
@sonney2kok16:41
-!- karlnapf [~heiko@host86-181-81-215.range86-181.btcentralplus.com] has left #shogun []16:41
wikingdorkmork... rebase heeellll16:42
-!- nickon [~noneedtok@kotnet-147.kulnet.kuleuven.be] has quit [Ping timeout: 265 seconds]16:49
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking]16:52
@sonney2kwiking don't rebase16:54
@sonney2kuse git format-patch16:54
n4nd0sonney2k: I am going to open soon a pull request with the so stuff I've done so far, do you think it could be possible to start merging some things?17:05
n4nd0I turned out not to be so straightforward to apply the patches this morning :(17:05
n4nd0some of them failed to apply17:05
-!- gsomix [~gsomix@188.168.2.219] has joined #shogun17:08
gsomixhi all17:08
gsomixpassed two pass-fail exams17:09
n4nd0gsomix: congratulations!17:11
n4nd0gtg, cu17:12
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving]17:12
-!- wiking [~wiking@2002:4e17:bd70:1:61bc:b9c1:1954:d673] has joined #shogun17:24
-!- wiking [~wiking@2002:4e17:bd70:1:61bc:b9c1:1954:d673] has quit [Changing host]17:24
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun17:24
-!- uricamic [~uricamic@2001:718:2:1634:fda6:1b8d:2f96:ced1] has quit [Quit: Leaving.]17:28
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun17:31
-!- pluskid [~pluskid@li400-235.members.linode.com] has quit [Quit: Leaving]17:40
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun17:44
-!- nickon [~noneedtok@d54C1F8A8.access.telenet.be] has joined #shogun17:48
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Ping timeout: 245 seconds]17:58
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun18:00
blackburnsonney2k: around?18:15
CIA-113shogun: Sergey Lisitsyn master * r7bbda1c / src/shogun/multiclass/MulticlassLibLinear.cpp : Fixed length of w vector in multiclass liblinaer - http://git.io/HyDcFw18:17
blackburnhmm I forgot what I wanted to ask18:17
CIA-113shogun: Sergey Lisitsyn master * r495a859 / (4 files in 2 dirs): Moved train to train_machine in AveragedPerceptron and OnlineLibLinear - http://git.io/96N12g19:00
-!- blackburn [~blackburn@188.122.250.167] has quit [Remote host closed the connection]20:41
-!- blackburn [~blackburn@188.122.250.167] has joined #shogun20:41
shogun-buildbotbuild #238 of nightly_none is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/23821:09
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Ping timeout: 245 seconds]21:12
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has joined #shogun21:29
shogun-buildbotbuild #247 of nightly_all is complete: Success [build successful]  Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_all/builds/24721:29
-!- puffin444 [62e3926e@gateway/web/freenode/ip.98.227.146.110] has quit [Ping timeout: 245 seconds]21:35
-!- gsomix [~gsomix@188.168.2.219] has quit [Quit: Ex-Chat]22:33
-!- nickon [~noneedtok@d54C1F8A8.access.telenet.be] has quit [Read error: Connection reset by peer]23:38
-!- blackburn [~blackburn@188.122.250.167] has left #shogun []23:59
--- Log closed Thu May 24 00:00:41 2012

Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!