--- Log opened Thu Apr 19 00:00:19 2012 | ||
blackburn | n4nd0: and yours? | 00:00 |
---|---|---|
n4nd0 | blackburn: databases, microwave engineering and wireless communications (dsp for wireless communications) | 00:01 |
n4nd0 | what do you study on dsp? | 00:01 |
blackburn | n4nd0: FFT, random signals processing | 00:03 |
n4nd0 | I think that is kind of cool :) | 00:03 |
blackburn | that's always better than optics | 00:04 |
blackburn | (actually everything is better than optics) | 00:04 |
blackburn | to eat some dirt for example is way much better than do some optics I think :D | 00:04 |
blackburn | n4nd0: ok see you | 00:17 |
n4nd0 | blackburn: good night! | 00:17 |
blackburn | good night | 00:17 |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed] | 00:17 | |
-!- blackburn [~qdrgsm@188.168.2.179] has left #shogun [] | 00:17 | |
-!- emrecelikten [~emre@92.44.168.7] has joined #shogun | 00:29 | |
-!- av3ngr [av3ngr@nat/redhat/x-jczscjkexunpkdrx] has joined #shogun | 01:39 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 260 seconds] | 01:51 | |
-!- emrecelikten1 [~emre@176.40.238.20] has joined #shogun | 02:10 | |
-!- emrecelikten [~emre@92.44.168.7] has quit [Ping timeout: 276 seconds] | 02:11 | |
-!- emrecelikten1 [~emre@176.40.238.20] has quit [Quit: Leaving.] | 02:18 | |
-!- PhilTillet [~Philippe@157.159.42.154] has joined #shogun | 02:20 | |
-!- PhilTillet [~Philippe@157.159.42.154] has quit [Read error: Connection reset by peer] | 02:37 | |
-!- vikram360 [~vikram360@117.192.163.173] has quit [Ping timeout: 252 seconds] | 03:01 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 07:35 | |
CIA-64 | shogun: Sergey Lisitsyn master * rad2707e / (35 files in 4 dirs): Moved some libraries to separate folder - http://git.io/YIJTHw | 07:57 |
CIA-64 | shogun: Soeren Sonnenburg master * r48b0dd1 / (35 files in 4 dirs): | 07:57 |
CIA-64 | shogun: Merge pull request #454 from lisitsyn/external_libs | 07:57 |
CIA-64 | shogun: move a number of classes not to be used from the outside to lib/external - http://git.io/J0NSow | 07:57 |
shogun-buildbot | build #738 of libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/738 blamelist: blackburn91@gmail.com | 08:02 |
CIA-64 | shogun: Soeren Sonnenburg master * r12a09b4 / (2 files in 2 dirs): move graphical regression_lars example to graphical - http://git.io/h6jxSA | 08:03 |
CIA-64 | shogun: Soeren Sonnenburg master * r677d721 / (2 files): fix wrong include paths for libsvm/liblinear - http://git.io/3Jhkyg | 08:10 |
@sonney2k | blackburn - hey look the new version of liblinear arrived http://www.csie.ntu.edu.tw/~cjlin/liblinear/ | 08:11 |
n4nd0 | sonney2k: goood morning! | 08:12 |
n4nd0 | I saw a bunch of new warning caused by cover tree stuff, mainly methods, args, etc not documented | 08:12 |
n4nd0 | I will try to fix them asap | 08:12 |
@sonney2k | n4nd0, thanks | 08:13 |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 09:12 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 09:12 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 09:12 | |
CIA-64 | shogun: Soeren Sonnenburg master * rc58a034 / src/shogun/classifier/svm/LibLinear.cpp : fix tron path in liblinear - http://git.io/x2shiw | 09:16 |
CIA-64 | shogun: Soeren Sonnenburg master * r519d74a / (4 files in 4 dirs): fix various documentations and thus doxygen warnings - http://git.io/wEzAJA | 09:20 |
shogun-buildbot | build #742 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/742 | 09:20 |
wiking | yo | 09:43 |
wiking | i have a problem with the HEAD | 09:43 |
wiking | ../shogun/lib/JLCoverTree.h:11:10: fatal error: 'values.h' file not found | 09:43 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 09:49 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 09:57 | |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 10:02 | |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun | 10:09 | |
sonne|work | wiking - this file does not exist - run git clean -dfx to clean up the mess :) | 10:14 |
shogun-buildbot | build #497 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/497 | 10:22 |
-!- PhilTillet [~android@92.90.16.93] has joined #shogun | 10:36 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 10:56 | |
blackburn | sonne|work: wow why did you merge it? | 10:57 |
blackburn | (I planned to complete it) | 10:58 |
blackburn | gsomix: n4nd0: any of you around? | 10:58 |
sonne|work | blackburn: I thought you did | 10:59 |
sonne|work | at least there was no indication that you didn't | 10:59 |
blackburn | sonne|work: ok sorry I'll mention that next time :) | 10:59 |
blackburn | some stuff to go still | 10:59 |
blackburn | ssl.cpp etc | 11:00 |
sonne|work | blackburn: btw have you seen liblinear :) | 11:00 |
sonne|work | it has regression now | 11:00 |
blackburn | sonne|work: yes checking right now | 11:00 |
blackburn | sonne|work: do you think it is so useful? | 11:00 |
sonne|work | I plan to merge it today | 11:00 |
sonne|work | yes of course | 11:00 |
sonne|work | svr is pretty good | 11:00 |
blackburn | sonne|work: will you? | 11:01 |
sonne|work | it should be easy | 11:01 |
blackburn | I mean I can | 11:01 |
blackburn | if it is needed :) | 11:01 |
sonne|work | after all SVR is reduced to svm prolbme | 11:01 |
n4nd0 | blackburn: yeah I am here | 11:01 |
sonne|work | you have exams :D | 11:01 |
sonne|work | so no time | 11:01 |
sonne|work | but I have a train ride :) | 11:01 |
blackburn | sonne|work: yes I got 5 today for my database project | 11:01 |
sonne|work | good luck then | 11:02 |
blackburn | sonne|work: damned optics tomorrow | 11:02 |
blackburn | n4nd0: do you need more tasks? | 11:03 |
blackburn | or some exam holidays? ;) | 11:03 |
n4nd0 | blackburn: mmm it sounds appealing but I have still stuff to finish in shogun and exams coming next week | 11:03 |
n4nd0 | blackburn: but tell me in any case, it is probably interesting | 11:04 |
blackburn | n4nd0: I have nothing in mind yet but if it is needed I'll come up with something ;) | 11:04 |
blackburn | sonne|work: can you guide me a little how to patch libocas to solve that | 11:05 |
blackburn | ||w||^2 + w0 w problem | 11:05 |
blackburn | ? | 11:05 |
sonne|work | blackburn: and n4nd0 should do this cool benchmarking problem :) | 11:07 |
sonne|work | blackburn: ask vojtech directly | 11:07 |
blackburn | sonne|work: covertree / qsort? | 11:08 |
blackburn | sonne|work: ok I will | 11:08 |
sonne|work | blackburn: faster way to compute distance | 11:08 |
blackburn | sonne|work: upper bounded? | 11:08 |
blackburn | btw I have seen harshit was talking about huge dataset yesterday | 11:09 |
n4nd0 | yeah | 11:09 |
n4nd0 | he wants to try knn there | 11:09 |
blackburn | it was 4000 vectors? | 11:09 |
n4nd0 | I understood that dimension of the vectors was 4000 | 11:09 |
blackburn | ghm what a huge dataset ;) | 11:09 |
sonne|work | blackburn: no by doing ||x-y||^2 = ||x||^2 -2xy + ||y||^2 | 11:10 |
-!- av3ngr [av3ngr@nat/redhat/x-jczscjkexunpkdrx] has quit [Quit: That's all folks!] | 11:10 | |
blackburn | sonne|work: precompute xy? | 11:10 |
sonne|work | x^Tx can be precomputed | 11:10 |
n4nd0 | the tests I am doing around every day with qsort / covertree are already bigger than 4000 vectors ;) | 11:10 |
sonne|work | only x^T y needs to be done | 11:10 |
sonne|work | blackburn: that is what we use in gaussian kernel btw | 11:10 |
blackburn | sonne|work: yes I know | 11:10 |
sonne|work | quite a bit faster there I would assume | 11:10 |
blackburn | sonne|work: hmmm if we assume vectors as normalized | 11:11 |
blackburn | we can use only dot there | 11:11 |
blackburn | sounds like a good option to use for me | 11:11 |
n4nd0 | sonne|work, blackburn so that decomposition should make it faster than using cblas function for the norm? | 11:12 |
blackburn | n4nd0: hmmmm | 11:13 |
-!- PhilTillet [~android@92.90.16.93] has quit [Read error: Connection reset by peer] | 11:13 | |
sonne|work | n4nd0: a lot - because you save even computing x-y | 11:14 |
blackburn | no need to compute square | 11:15 |
sonne|work | square is not the problem - it is done once at the end | 11:16 |
sonne|work | sqrt I meant | 11:16 |
sonne|work | forget what I said | 11:17 |
n4nd0 | ok ... | 11:17 |
blackburn | sonne|work: sqrt is not but (x-y)(x-y) ;) | 11:22 |
-!- PhilTillet [~android@92.90.16.93] has joined #shogun | 11:22 | |
-!- wiking [~wiking@iridia-pc21.ulb.ac.be] has joined #shogun | 11:22 | |
-!- wiking [~wiking@iridia-pc21.ulb.ac.be] has quit [Changing host] | 11:22 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 11:22 | |
sonne|work | yes | 11:22 |
sonne|work | wiking: see the very above | 11:23 |
wiking | ok i have a fix | 11:23 |
sonne|work | git clean -dfx | 11:23 |
wiking | mmm i've just got connected | 11:23 |
sonne|work | (it will erase everything not in git from shogun) | 11:23 |
sonne|work | wiking: so no need for a fix :D | 11:24 |
wiking | jlcovertree is not part of shogun? | 11:24 |
wiking | sonne|work: https://github.com/vigsterkr/shogun/commit/edd4b29fa0b09226c2bb34e99d601d6e2247fa40 | 11:25 |
wiking | otherwise it won't compile on a BSD-like system | 11:25 |
wiking | as it does not have <values.h> | 11:25 |
sonne|work | wiking: I see | 11:26 |
sonne|work | what is values used for at all? | 11:26 |
sonne|work | I mean why do we need it? | 11:26 |
sonne|work | n4nd0: any idea? | 11:26 |
wiking | hehehe no idea :) | 11:26 |
wiking | i've just sent you now the pull request | 11:26 |
sonne|work | does it compile if you remove it? | 11:26 |
n4nd0 | sonne|work: I am checking | 11:26 |
blackburn | what is values? | 11:27 |
blackburn | http://tigcc.ticalc.org/doc/values.html | 11:28 |
blackburn | some crazy consts | 11:28 |
n4nd0 | it seems that the file values.h is nowhere | 11:28 |
blackburn | mystery file | 11:29 |
wiking | blackburn: yo | 11:29 |
blackburn | wiking: hi | 11:29 |
wiking | blackburn: what was so funny yesterday? | 11:29 |
wiking | :D | 11:29 |
blackburn | wiking: your git hotline | 11:30 |
n4nd0 | I don't really know why my compilation didn't fail locally | 11:30 |
wiking | blackburn: well you it's a bit funny ;) | 11:30 |
sonne|work | n4nd0: linux | 11:30 |
sonne|work | I think we should modify the code to work w/o vlaues | 11:30 |
sonne|work | values.h | 11:30 |
wiking | n4nd0: as said you have linux | 11:30 |
wiking | sonne|work: or move values include under some base/ header | 11:30 |
blackburn | Hello, Git consulting, Viktor is on line | 11:31 |
wiking | hahahah yeah | 11:31 |
sonne|work | no | 11:31 |
wiking | and to see how fucken nice i am | 11:31 |
wiking | or unbiased | 11:31 |
wiking | i'm even supporting my competition | 11:31 |
wiking | ;) | 11:31 |
blackburn | olympics spirit lol | 11:31 |
wiking | yeps | 11:31 |
sonne|work | wiking: I don't think anyone here wants your project :D | 11:31 |
blackburn | sonne|work: huh! pull request is here | 11:31 |
wiking | sonne|work: yes yes they want! | 11:32 |
wiking | ok i have a lil libqp fix as well | 11:32 |
wiking | let me add the qp branch to my repo | 11:32 |
sonne|work | these bastards | 11:32 |
wiking | oh these mother!@# | 11:34 |
blackburn | sonne|work: I also thought nobody wants my project | 11:34 |
wiking | so i've added libqp_gsmo https://github.com/vigsterkr/shogun/commit/8e24e152bfba02cfc4e0890c485667455cbfcaf2 | 11:34 |
wiking | since it's included in libqp.h | 11:34 |
wiking | but the problem now that i cannot cherry-pick the pull request | 11:35 |
wiking | ;) | 11:35 |
blackburn | where did you get that? | 11:35 |
blackburn | ;) | 11:35 |
wiking | which what? | 11:35 |
wiking | libqp_gsm? | 11:35 |
wiking | gsmo | 11:35 |
wiking | it's from libqp | 11:35 |
sonne|work | wiking: it seems to compile here w/o values.h | 11:35 |
wiking | but it has never been added to shogun | 11:36 |
sonne|work | wiking: can you try if you need it at all under osx? | 11:36 |
wiking | sonne|work: ok just a sec | 11:36 |
wiking | n4nd0: so where did u get this code? | 11:37 |
wiking | :))) | 11:37 |
wiking | or why did you do the include <values.h> ? | 11:37 |
blackburn | can somebody send me an sms if chris arrives? :D | 11:38 |
n4nd0 | wiking: from there http://hunch.net/~jl/projects/cover_tree/cover_tree.html | 11:38 |
wiking | ah ok | 11:38 |
wiking | i wonder if it uses really any of those macros | 11:39 |
n4nd0 | I don't even know what is defined in values.h | 11:39 |
wiking | :D | 11:39 |
wiking | mm i think it compiles for me as well... although i'm still waiting for finishing it | 11:40 |
wiking | ok yeah now it's already doing doxygen | 11:40 |
wiking | so it was ok | 11:40 |
sonne|work | wiking: ok then | 11:40 |
wiking | sonne|work: yep compiles | 11:40 |
wiking | sonne|work: can u remove then the header | 11:40 |
wiking | and i'll merge my repo? | 11:40 |
sonne|work | letst remove the header | 11:40 |
CIA-64 | shogun: Soeren Sonnenburg master * r000a882 / src/shogun/lib/JLCoverTree.h : remove values.h include - http://git.io/iXGNbA | 11:41 |
wiking | ok cool | 11:41 |
wiking | i'll try to merge now | 11:41 |
wiking | ok here's my pull request for libqp | 11:46 |
wiking | don't worry about the 3 commits... it only adds the file itself.. i had to do a merge in order to not to fuck up totally my repo | 11:46 |
shogun-buildbot | build #744 of libshogun is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/744 blamelist: sonne@debian.org | 11:50 |
wiking | ahahahah | 11:50 |
wiking | MAXFLOAT | 11:50 |
wiking | :D | 11:50 |
sonne|work | wiking: looks like you forgot the rebase https://github.com/shogun-toolbox/shogun/pull/470 | 11:50 |
wiking | so fuck | 11:51 |
wiking | isn't maxfloat defined somewhere in shogun already? | 11:51 |
sonne|work | grep MAXFLOAT /usr/include/math.h | 11:52 |
wiking | noup | 11:52 |
wiking | not on my system | 11:52 |
wiking | no ok it'll work as well here | 11:53 |
sonne|work | # ifdef __USE_XOPEN | 11:53 |
sonne|work | /* X/Open wants another strange constant. */ | 11:53 |
sonne|work | # define MAXFLOAT 3.40282347e+38F | 11:53 |
sonne|work | # endif | 11:53 |
sonne|work | here hmmhh | 11:53 |
wiking | it's just in another header that it's included in math.h | 11:53 |
sonne|work | yeah but math is already included in that tree?! | 11:54 |
-!- al___ [4eecc9c5@gateway/web/freenode/ip.78.236.201.197] has joined #shogun | 11:58 | |
al___ | hi there | 11:58 |
al___ | compilation issue: | 11:58 |
al___ | In file included from classifier/KNN.cpp:18:0: ../shogun/lib/JLCoverTree.h: In function 'node<P> batch_create(shogun::v_array<T>)': ../shogun/lib/JLCoverTree.h:277:52: error: 'MAXFLOAT' was not declared in this scope ../shogun/lib/JLCoverTree.h: In function 'void batch_nearest_neighbor(const node<P>&, const node<P>&, shogun::v_array<shogun::v_array<T> >&)': ../shogun/lib/JLCoverTree.h:764:22: error: 'MAXFLOAT' was not | 11:58 |
al___ | with lastest git code | 11:58 |
al___ | thanks for helping | 11:59 |
blackburn | hahah guys that is really fast bugreport | 11:59 |
n4nd0 | indeed | 11:59 |
n4nd0 | so it looks like values.h was used for something :S | 12:00 |
al___ | :p indeed | 12:00 |
blackburn | n4nd0: just replace with CMath::ALMOST_INFTY or so? | 12:00 |
al___ | will try it right now | 12:00 |
n4nd0 | blackburn: I don't have access right now to the code | 12:00 |
n4nd0 | blackburn: I am not in my workstation | 12:01 |
blackburn | I see | 12:01 |
blackburn | al___: yeah you may try as well :) | 12:01 |
al___ | ../shogun/lib/JLCoverTree.h line 277 764 766 remplace MAXFLOAT with CMath::ALMOST_INFTY do the work | 12:05 |
al___ | thanks | 12:05 |
n4nd0 | al___: you can do a patch with that there ;) | 12:05 |
al___ | It will be a pleasure, can you tell me how to do it ? :p | 12:06 |
n4nd0 | you can issue a pull request from github | 12:08 |
blackburn | or create a patch and send it | 12:09 |
blackburn | sonne|work: can you apply a patch? | 12:09 |
blackburn | however PR is more preferrable I think | 12:10 |
wiking | :>> | 12:14 |
al___ | I m little bit nex to github, I can read the open PR but how can I create a new one ? | 12:19 |
wiking | :))) | 12:19 |
wiking | sonne|work: you should have just taken my patch :>> | 12:19 |
blackburn | al___: fork shogun - clone your fork - do some changes locally - push to your remote on github - create pull request using github interface | 12:21 |
wiking | fuuuucking hell my shogun fork is fsck-ed | 12:26 |
wiking | :> | 12:26 |
al___ | blackburn: I got a RSA permission denied when cloning my fork shogun | 12:32 |
n4nd0 | al___: are you following any tuto in github? | 12:32 |
blackburn | http://help.github.com/ | 12:33 |
blackburn | try to set up your git according to that link ^ | 12:33 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Remote host closed the connection] | 12:39 | |
-!- wiking [~wiking@vpnc027.ugent.be] has joined #shogun | 12:39 | |
-!- wiking [~wiking@vpnc027.ugent.be] has quit [Changing host] | 12:39 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 12:39 | |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 240 seconds] | 12:41 | |
wiking | ok i think i have the solution | 12:46 |
wiking | just somebody help me to rebase my giiit! | 12:46 |
n4nd0 | wiking: are you kidding? | 12:48 |
wiking | yep no | 12:48 |
n4nd0 | I don't have the source with me right now, sorry :( | 12:48 |
wiking | n4nd0: no worries | 12:49 |
wiking | i think we should use FLT_MAX | 12:49 |
wiking | since it's iso c standard | 12:49 |
wiking | not like MAXFLOAT | 12:49 |
wiking | that is xsi | 12:49 |
wiking | anybody here on debian or has access to debian | 12:49 |
wiking | that can test this for me? | 12:49 |
wiking | although according to the repo FLT_MAX has been used in other sources and since those compiles... | 12:50 |
blackburn | wiking: why not CMath's constant? | 12:53 |
wiking | because it's not really the flt_max | 12:53 |
blackburn | why not to change CMath's stuff to real max? | 12:54 |
wiking | well it's not the same | 12:54 |
wiking | :) | 12:54 |
blackburn | actually I am not used to work with e+38 distances and hardly can imagine that :) | 12:55 |
wiking | well | 12:55 |
wiking | afaik it's there because to check float overflow | 12:56 |
blackburn | hmm I thought it is kind of maximal distance | 12:56 |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 12:58 | |
blackburn | drum roll chris is arriving! | 12:58 |
wiking | ok anyhow FLT_MAX will do the work | 12:58 |
blackburn | :D | 12:59 |
-!- ckwidmer [~chris@HSI-KBW-046-005-237-106.hsi8.kabel-badenwuerttemberg.de] has joined #shogun | 12:59 | |
blackburn | tadaam | 12:59 |
wiking | \o/ | 12:59 |
wiking | blackburn: pull request is there | 13:01 |
blackburn | wiking: I am not able to merge it right now anyway ;) | 13:01 |
blackburn | will do in a hour once I get home | 13:02 |
wiking | okey | 13:02 |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 260 seconds] | 13:09 | |
CIA-64 | shogun: Soeren Sonnenburg master * r2132e4e / src/shogun/lib/JLCoverTree.h : change MAXFLOAT -> FLT_MAX - http://git.io/O0lrnw | 13:11 |
wiking | yeey | 13:11 |
sonne|work | wiking: it is soo easy to make you happy :D | 13:11 |
wiking | although | 13:11 |
wiking | should have taken my pull request | 13:11 |
wiking | because this way now we have this | 13:11 |
wiking | https://github.com/shogun-toolbox/shogun/pull/470 | 13:11 |
wiking | :DDD | 13:11 |
sonne|work | wiking: sry didn't see this | 13:11 |
wiking | ok i'll try now a rebase :DDD | 13:12 |
wiking | it'll be horrible for sure | 13:12 |
wiking | ok wasn't that bad as i thought it would be | 13:12 |
sonne|work | wiking: I guess it is about time to play with git reset --hard :D | 13:14 |
shogun-buildbot | build #745 of libshogun is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/libshogun/builds/745 | 13:14 |
sonne|work | huray | 13:15 |
wiking | sonne|work: yeps | 13:16 |
sonne|work | wiking about libqp - shouldn't you rather incorporate / take the whole beast at once? | 13:16 |
wiking | sonne|work: ? | 13:16 |
sonne|work | whole lipqp? | 13:16 |
wiking | you mean when i already have the wrappers? | 13:16 |
wiking | i mean otherwise the whole libqp is there | 13:16 |
wiking | i've checked the latest code of libqp | 13:16 |
wiking | the _splx in shogun was up to date | 13:16 |
wiking | and the header as well | 13:16 |
sonne|work | ahh so the rest is up to date? | 13:17 |
sonne|work | then ok | 13:17 |
wiking | yep | 13:17 |
wiking | let me just fix again my repo | 13:17 |
wiking | :)) | 13:17 |
sonne|work | wiking: https://github.com/shogun-toolbox/shogun/pull/470/commits | 13:17 |
sonne|work | yeah one request only please | 13:17 |
wiking | yep | 13:17 |
wiking | i'm just trying to fix it now | 13:17 |
sonne|work | so git reset --hard should do and a pull with a rebase and push --force | 13:17 |
wiking | ok | 13:18 |
wiking | rebased | 13:18 |
wiking | check now the link | 13:18 |
wiking | it's only 1 commit now | 13:18 |
CIA-64 | shogun: Viktor Gal master * rfc8fa5e / src/shogun/lib/external/libqp_gsmo.cpp : | 13:20 |
CIA-64 | shogun: Add libqp_gsmo implementation from libqp | 13:20 |
CIA-64 | shogun: As libqp.h contains the definition for libqp_gsmo_solver | 13:20 |
CIA-64 | shogun: the implementation of it should be included as well in shogun - http://git.io/sZObSA | 13:20 |
CIA-64 | shogun: Soeren Sonnenburg master * r582e0cc / src/shogun/lib/external/libqp_gsmo.cpp : | 13:20 |
CIA-64 | shogun: Merge pull request #470 from vigsterkr/qp | 13:20 |
CIA-64 | shogun: add libqp_gsmo_solver - http://git.io/_6v3Rw | 13:20 |
sonne|work | there we go | 13:20 |
sonne|work | now we should get pr_loqo.cpp (in shogun already) into shape to match libqp's interface | 13:21 |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 13:21 | |
wiking | mmmm | 13:23 |
wiking | loll good that you are telling me this | 13:23 |
wiking | :) | 13:23 |
sonne|work | and actually while we are at it also CTron | 13:23 |
blackburn | sonne|work: I was like O_O when I saw predictor-corrector stuff in pr_loqo | 13:23 |
sonne|work | ?! | 13:25 |
n4nd0 | wiking: so is it libq integration in shogun better now? | 13:26 |
wiking | n4nd0: not even near to it | 13:26 |
n4nd0 | wiking: lot of stuff left to do still? | 13:26 |
wiking | yeah well we need a descent c++ wrapper for all the qp solvers | 13:27 |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 252 seconds] | 13:27 | |
sonne|work | n4nd0: currently code got just copied in | 13:28 |
sonne|work | no mor | 13:28 |
sonne|work | e | 13:28 |
n4nd0 | I see | 13:29 |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 13:30 | |
wiking | sonne|work: so by the way... is there any other qp solvers in shogun? | 13:30 |
wiking | and shouldn't we move qp solvers out of classifiers? | 13:31 |
sonne|work | wiking: that is what I said :) | 13:32 |
wiking | hahahah where? :) | 13:32 |
sonne|work | well pr_loqo and tron | 13:33 |
sonne|work | but IMHO these should go into lipqp | 13:33 |
wiking | into? | 13:33 |
wiking | in what sense | 13:33 |
wiking | i mean libqp is just as is | 13:33 |
wiking | let's separate libqp and libqp wrapping | 13:34 |
wiking | imho | 13:34 |
wiking | i suppose you are talking about now libqp wrapper functions | 13:34 |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 248 seconds] | 13:36 | |
-!- ckwidmer [~chris@HSI-KBW-046-005-237-106.hsi8.kabel-badenwuerttemberg.de] has quit [Remote host closed the connection] | 13:41 | |
blackburn | drum roll again | 13:42 |
blackburn | chris is leaving | 13:42 |
blackburn | and I am going too as well | 13:42 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 13:42 | |
-!- al___ [4eecc9c5@gateway/web/freenode/ip.78.236.201.197] has quit [Quit: Page closed] | 13:48 | |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 13:54 | |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 276 seconds] | 14:01 | |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed] | 14:03 | |
sonne|work | wiking: no I am talking about making it libqp_prloqo.cpp | 14:04 |
wiking | ? | 14:04 |
sonne|work | and getting it core libqp | 14:04 |
sonne|work | and on top of it the qp interface | 14:05 |
wiking | so you wanna put prloqo into libqp? | 14:05 |
sonne|work | yes | 14:05 |
wiking | or make it as part of it | 14:05 |
wiking | ah ok | 14:05 |
sonne|work | and tron | 14:05 |
wiking | tron as well | 14:05 |
wiking | cool | 14:05 |
wiking | can we move them to maybe | 14:05 |
wiking | lets say src/shogun/qp | 14:05 |
wiking | ? | 14:05 |
wiking | or something like that? | 14:05 |
wiking | as it's not an external lib | 14:07 |
sonne|work | wiking: I thought so too but now that blackburn started this lib/external stuff I think these should be in there ... maybe external/libqp | 14:07 |
wiking | mmm but they are not part of libqp | 14:07 |
wiking | so i mean when let's say | 14:07 |
wiking | vojtech updates the libqp | 14:07 |
sonne|work | no but they should become part of it | 14:07 |
wiking | it's pretty easy to do an upgrade | 14:07 |
sonne|work | let me talk to vojech | 14:07 |
wiking | ok | 14:07 |
sonne|work | about this | 14:07 |
wiking | talk with him decide and then i'll move it | 14:07 |
wiking | but afaik we'll still need a src/shogun/qp | 14:07 |
sonne|work | yes | 14:07 |
wiking | to put there the wrapper classes | 14:08 |
sonne|work | maybe better shogun/optimizer | 14:08 |
wiking | ok | 14:08 |
sonne|work | and have the qp stuff in there | 14:08 |
sonne|work | and some linear stuff etc too | 14:08 |
wiking | that sounds good | 14:08 |
wiking | ok i'll then make an optimizer dir and put there the abstract wrapper as a starter | 14:08 |
wiking | and then we'll see what to do with pr_loqo and tron | 14:08 |
sonne|work | have a look at cplex.h too maybe remotely kind of this | 14:09 |
wiking | mmmhm | 14:09 |
wiking | okey | 14:09 |
wiking | but first things first i want to finish my simple cccp implementation for latent | 14:10 |
wiking | today i've done already too much branch switching till now :) | 14:10 |
-!- blackburn [~qdrgsm@188.168.2.179] has joined #shogun | 14:14 | |
-!- xiangwang [~chatzilla@159.226.60.224] has joined #shogun | 14:17 | |
sonne|work | wiking: as you wish :) | 14:19 |
-!- PhilTillet [~android@92.90.16.93] has quit [Ping timeout: 240 seconds] | 14:21 | |
-!- xiangwang [~chatzilla@159.226.60.224] has quit [Ping timeout: 252 seconds] | 14:25 | |
wiking | sonne|work: let me know when you've concluded the business with vojtech ;) | 14:26 |
sonne|work | :P | 14:26 |
-!- gsomix [~gsomix@188.168.2.191] has quit [Ping timeout: 260 seconds] | 14:37 | |
-!- pluskid [~chatzilla@111.120.4.214] has joined #shogun | 14:51 | |
sonne|work | hey pluskid | 14:51 |
sonne|work | welcome back | 14:51 |
pluskid | :) | 14:51 |
pluskid | just back from the trip | 14:51 |
pluskid | I saw your feedbacks | 14:51 |
sonne|work | I guessed so | 14:51 |
pluskid | will fix those soon (tomorrow) | 14:52 |
pluskid | too tired to work tonight | 14:52 |
sonne|work | no worries | 14:52 |
sonne|work | if none of you applied at a different org we already now know who will be in | 14:52 |
pluskid | btw, do you think I should keep the code for reading and writing multiclass strategy in MulticlassSVM? | 14:52 |
pluskid | I only applied shogun | 14:53 |
pluskid | so am I in? :p | 14:53 |
sonne|work | hehe | 14:53 |
pluskid | won't tell me this moment? haha, bad boy | 14:53 |
sonne|work | that only means you are selected by us then you will be in :) | 14:53 |
sonne|work | I am not allowed to | 14:53 |
sonne|work | but I don't think it is too hard to guess | 14:54 |
pluskid | btw, I'm just curious, sonne|work , do you hate STL? :p | 14:54 |
sonne|work | anyways multiclass svm -> maybe multiclass kernel machine? | 14:54 |
sonne|work | pluskid: yes but don't ask my why | 14:54 |
sonne|work | blackburn asked me that so often that I forgot | 14:54 |
sonne|work | except that it looks ugly (syntax wise) | 14:55 |
wiking | when is the list going to be released? | 14:55 |
pluskid | OK, that's why they decided to sent you a gift of STL book, haha | 14:55 |
wiking | only next tuesday? | 14:55 |
sonne|work | google will announce on 23rd | 14:55 |
pluskid | I can also feel your hate toward STL, too :) | 14:55 |
wiking | STL for sonne|work \o/ | 14:55 |
sonne|work | *grrgg* | 14:55 |
wiking | but honestly one should consider valarray :) | 14:55 |
sonne|work | wiking: we have SGVector :P | 14:56 |
pluskid | I think the read/write of multiclass-strategy should be put in a read/write function in MulticlassMachine | 14:56 |
wiking | sonne|work: yes but within sgvector | 14:56 |
pluskid | but it seems current there's no such function in MulticlassMachine | 14:56 |
wiking | instead of having float* | 14:56 |
sonne|work | pluskid: multiclasskernelmachine | 14:56 |
wiking | a valarray would allow a lot of automatic optimization... | 14:56 |
sonne|work | these functions are not used anyways I would guess | 14:56 |
pluskid | why? | 14:56 |
sonne|work | maybe some used them long time ago for multiclass svms | 14:56 |
pluskid | hmm | 14:57 |
pluskid | then I'll not worry too much about that | 14:57 |
pluskid | I'd try to keep it similar to the original code | 14:57 |
sonne|work | but nowadays the save_serializable / load_serializable stuff is much more powerful/usefull | 14:57 |
pluskid | so, this two manually crafted function might be removed some day | 14:58 |
sonne|work | I think all load/save() functions should go | 14:58 |
pluskid | wiking: what is a valarray? | 14:58 |
pluskid | sonne|work: at this moment? | 14:58 |
sonne|work | no | 14:58 |
wiking | pluskid: the forgotten child of STL | 14:58 |
wiking | http://www.cplusplus.com/reference/std/valarray/ | 14:59 |
sonne|work | wiking: why the massive overhead? | 14:59 |
wiking | it's really a very very small overhead | 14:59 |
wiking | compared to other STL containers | 14:59 |
wiking | as this is really not a real STL container | 14:59 |
sonne|work | hehe | 14:59 |
sonne|work | we simply want to be fast for low-dimensional data too | 15:00 |
wiking | and recent g++ implementation use a done a lot of optimization with that | 15:00 |
wiking | i.e. using SIMD and multiple thread out of box | 15:00 |
sonne|work | and having more than size / refcount as overhead is imho not acceptable | 15:00 |
wiking | since valarray is a really simple class | 15:00 |
pluskid | oh, it seems what a valarray is what we wish to do with our SGVector? | 15:00 |
sonne|work | pluskid: yeah the a+b, sin(a) etc operations | 15:01 |
wiking | sonne|work: check out the definition of valarray template: /usr/include/c++/4.2.1/valarray | 15:01 |
wiking | private: | 15:02 |
wiking | size_t _M_size; | 15:02 |
wiking | _Tp* __restrict__ _M_data; | 15:02 |
wiking | 15:02 | |
wiking | this is all the overhead ;) | 15:02 |
pluskid | btw, I seriously think we need some basic matrix operations support after doing the LARS work, Blas/LaPack is a bit difficult to use, especially for simple tasks like dot product of vectors, we need some wrappers | 15:02 |
pluskid | wiking: maybe we need to do some standard benchmarks someday to persuade sonne|work that STL (at least valarray) is not that evil | 15:04 |
wiking | pluskid: hehehe there's already such available... just google it | 15:04 |
pluskid | wiking: about valarray or STL? | 15:04 |
wiking | valarray | 15:04 |
pluskid | I googled "valarray benchmark", and the first result it "why is valarray so slow", haha | 15:05 |
wiking | hahaha | 15:06 |
wiking | stackoverflow | 15:06 |
wiking | :) | 15:06 |
wiking | pluskid: http://www.iue.tuwien.ac.at/phd/heinzl/img760.png | 15:07 |
wiking | pluskid: http://www.iue.tuwien.ac.at/phd/heinzl/node47.html | 15:07 |
pluskid | cool, seems especially optimized for "small" (10^3) sized vector | 15:10 |
wiking | mm | 15:10 |
wiking | i mean in this sense maybe having GSL will solve all the problems ;) | 15:10 |
pluskid | haha, another huge dependency that I guess sonne|work also hates :p | 15:11 |
wiking | hahaha | 15:12 |
pluskid | be back in a while | 15:12 |
wiking | but gsl is just about over optimizing all kind of vector/matrix operations | 15:12 |
-!- pluskid_ [~chatzilla@1.204.114.149] has joined #shogun | 15:25 | |
-!- pluskid [~chatzilla@111.120.4.214] has quit [Read error: Connection reset by peer] | 15:25 | |
-!- pluskid_ is now known as pluskid | 15:25 | |
blackburn | hehe | 15:25 |
wiking | mmm | 15:26 |
wiking | anybody else might need a CMath:: function that not only returns the max value of a vector but the position/index of that max value as well? | 15:26 |
blackburn | wiking: isn't it available already? | 15:27 |
wiking | mmm is it? i was looking for ::max | 15:27 |
wiking | only found static inline T max(T* vec, int32_t len) | 15:27 |
wiking | till now | 15:27 |
blackburn | arg_max? | 15:27 |
wiking | noup | 15:28 |
wiking | there's no such | 15:28 |
wiking | or at least i couldn't find till now in Math.h | 15:29 |
blackburn | line 900 or so | 15:29 |
sonne|work | static inline int32_t arg_max(T * vec, int32_t inc, int32_t len, T * maxv_ptr = NULL) | 15:29 |
wiking | lol? | 15:30 |
wiking | git grep arg_max | 15:30 |
wiking | 0 resuls | 15:30 |
sonne|work | no objections for having a simpler one though | 15:30 |
wiking | results | 15:30 |
wiking | is it maybe in a recent commit? | 15:30 |
blackburn | yes | 15:30 |
blackburn | by fernando I guess | 15:30 |
sonne|work | QDA commit | 15:30 |
wiking | since i haven't rebased my latent branch for a hiwhile | 15:30 |
wiking | ah ok | 15:30 |
blackburn | shogun/classifier/QDA.cpp: out->set_label(i, CMath::arg_max(norm2.vector+i, num_vecs, m_num_classes)); | 15:30 |
blackburn | shogun/mathematics/Math.h: /// return arg_max(vec) | 15:30 |
blackburn | shogun/mathematics/Math.h: static inline int32_t arg_max(T * vec, int32_t inc, int32_t len, T * maxv_ptr = NULL) | 15:30 |
wiking | so that much be the reason | 15:31 |
wiking | ok i'll rebase then | 15:31 |
-!- gsomix [~gsomix@83.234.54.111] has joined #shogun | 15:33 | |
blackburn | sonne|work: chris is in tuebingen/berlin up to may so exclude the usa for kickoff | 15:34 |
sonne|work | except if we have a student there | 15:34 |
blackburn | sh | 15:35 |
blackburn | forget what I said then heh | 15:42 |
-!- blackburn [~qdrgsm@188.168.2.179] has quit [Ping timeout: 246 seconds] | 15:51 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 15:53 | |
-!- pluskid [~chatzilla@1.204.114.149] has quit [Ping timeout: 246 seconds] | 16:17 | |
-!- pluskid [~chatzilla@1.204.114.149] has joined #shogun | 16:22 | |
-!- gsomix [~gsomix@83.234.54.111] has quit [Ping timeout: 260 seconds] | 16:44 | |
-!- Marty28 [~marty@158.181.78.199] has joined #shogun | 16:47 | |
-!- gsomix [~gsomix@188.168.14.33] has joined #shogun | 16:59 | |
-!- Marty28 [~marty@158.181.78.199] has quit [Quit: Colloquy for iPad - http://colloquy.mobi] | 17:07 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Read error: Operation timed out] | 17:35 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun | 17:47 | |
-!- pluskid [~chatzilla@1.204.114.149] has quit [Ping timeout: 264 seconds] | 18:13 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 18:44 | |
wiking | eeey | 18:55 |
-!- wiking_ [~wiking@iridia-pc21.ulb.ac.be] has joined #shogun | 18:58 | |
-!- wiking_ [~wiking@iridia-pc21.ulb.ac.be] has quit [Changing host] | 18:58 | |
-!- wiking_ [~wiking@huwico/staff/wiking] has joined #shogun | 18:58 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 248 seconds] | 19:01 | |
-!- wiking_ is now known as wiking | 19:01 | |
* sonney2k alright | 19:07 | |
* sonney2k lets kick as and chew bubblegum | 19:08 | |
* sonney2k goes berserk and updates liblinear | 19:08 | |
wiking | mmm | 19:08 |
wiking | i'm gonna use ocas as a solver for now in latent svm | 19:08 |
wiking | that's what i've decided :) | 19:08 |
wiking | ok off to train | 19:18 |
wiking | ttyl | 19:18 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 19:19 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 264 seconds] | 19:29 | |
gsomix | it's seems that `sublime text 2' become my favourite editor | 19:38 |
gsomix | I still not able to use vim... | 19:39 |
@sonney2k | gsomix, try vimtutor | 19:39 |
gsomix | sonney2k, I tried :) | 19:40 |
CIA-64 | shogun: Soeren Sonnenburg master * r9815db6 / src/shogun/lib/external/libqp_gsmo.cpp : fix compile warning in libqp_gsmo - http://git.io/0U6Zlg | 19:47 |
CIA-64 | shogun: Soeren Sonnenburg master * r723ccd1 / (2 files in 2 dirs): fix doxygen documentation related errors - http://git.io/G4j8jA | 19:47 |
n4nd0 | gsomix: maybe you are an emacs person ;) | 20:03 |
gsomix | n4nd0, huh | 20:04 |
* sonney2k thinks that gsomix still needs his fingers | 20:11 | |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun | 20:16 | |
gsomix | n4nd0, which exams in this year? | 20:26 |
n4nd0 | gsomix: so I'll do next week databases and microwave engineering | 20:27 |
n4nd0 | gsomix: wireless communications in a couple of weeks | 20:27 |
gsomix | >> databases | theoretical course? | 20:28 |
n4nd0 | not really | 20:28 |
n4nd0 | the first semester was rather theoretical | 20:28 |
n4nd0 | now it is half of it sql | 20:28 |
n4nd0 | and other half security and normlization | 20:28 |
n4nd0 | normalization, | 20:28 |
n4nd0 | sorry | 20:28 |
n4nd0 | gsomix: what about you? | 20:29 |
n4nd0 | optics? | 20:30 |
gsomix | n4nd0, math analysis, complex analysis, philosophy and optics | 20:30 |
gsomix | in june, i think. | 20:30 |
gsomix | n4nd0, I have horrible course about databases... FoxPro 9 =____= | 20:31 |
gsomix | *in last year | 20:31 |
n4nd0 | FoxPro 9, no idea | 20:31 |
n4nd0 | I have never taken a course itself in complex analysis, but I have done stuff with transforms | 20:32 |
n4nd0 | and last year I used it quite a bit in a course in electromagnetism | 20:32 |
n4nd0 | I think complex analysis looks nice | 20:32 |
n4nd0 | do you like it? | 20:32 |
gsomix | n4nd0, yep. but i'm missing a lot of lectures about it. | 20:34 |
gsomix | >> in a course in electromagnetism | aha, impedances and other | 20:35 |
-!- puffin444 [230bf329@gateway/web/freenode/ip.35.11.243.41] has joined #shogun | 20:35 | |
n4nd0 | in this one impedances didn't appear that much | 20:36 |
n4nd0 | it was about electromagnetic propagation in plasma media | 20:36 |
n4nd0 | with tensors and stuff like that | 20:37 |
puffin444 | hello | 20:37 |
n4nd0 | hey puffin444 | 20:37 |
puffin444 | Have I walked in the middle of a physics conversation? :) | 20:37 |
gsomix | n4nd0, I see | 20:37 |
n4nd0 | haha just asking about our forthcoming exams | 20:37 |
puffin444 | Oh are you students of physics? | 20:37 |
n4nd0 | telecommunications + CS here :) | 20:38 |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 252 seconds] | 20:38 | |
gsomix | puffin444, I'm undergraduate student in physics. | 20:39 |
gsomix | puffin444, ah, hello :) | 20:39 |
puffin444 | Oh I see. I'm CS+Math+Econ | 20:39 |
puffin444 | Hi | 20:40 |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has joined #shogun | 20:41 | |
gsomix | what is english equivalent of exam without assessment? | 20:42 |
n4nd0 | what do you mean? | 20:42 |
puffin444 | An ungraded exam? | 20:43 |
@sonney2k | puh, liblinear 1.9 has *lots* of changes | 20:43 |
@sonney2k | too many for any train ride | 20:43 |
n4nd0 | sonney2k: I think you need the Siberian train for that :) | 20:45 |
gsomix | n4nd0, hmm, exams with only binary assessment (pass / fail) | 20:45 |
gsomix | proof that you attended a classes. | 20:47 |
gsomix | ok, never mind :) | 20:47 |
-!- PhilTillet [~Philippe@157.159.42.154] has joined #shogun | 20:47 | |
@sonney2k | n4nd0, heh I think some train to frankfurt would suffice | 20:48 |
@sonney2k | but still I didn't know how much (subtle) changes they did even to old liblinear | 20:48 |
@sonney2k | this + our modifications don't make live so easy | 20:49 |
puffin444 | How does liblinear fit in with shogun? | 20:49 |
PhilTillet | helo everybody :) | 20:49 |
gsomix | PhilTillet, hi | 20:51 |
puffin444 | PhilTillet, hey | 20:52 |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has quit [Quit: Page closed] | 20:52 | |
PhilTillet | what's up guys? | 20:52 |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has joined #shogun | 20:53 | |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has quit [Client Quit] | 20:55 | |
PhilTillet | I have powned my keyboard??? | 21:01 |
PhilTillet | oops | 21:01 |
PhilTillet | ...* | 21:01 |
PhilTillet | I have somewhat broken the "j" key | 21:01 |
puffin444 | What happened? You can't press the "j" key? | 21:02 |
PhilTillet | I can :p but it fell off my keyboard, and it is broken so I can't put it back :/ | 21:06 |
puffin444 | That's quite annoying. That happened to my touchpad mouse buttons. | 21:08 |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has joined #shogun | 21:08 | |
PhilTillet | XD | 21:10 |
PhilTillet | yes | 21:10 |
PhilTillet | I think i'll try to carefully glue it | 21:10 |
-!- PhilTillet [~Philippe@157.159.42.154] has quit [Quit: Leaving] | 21:14 | |
-!- PhilTillet [~Philippe@157.159.42.154] has joined #shogun | 21:14 | |
PhilTillet | sonney2k, I still can't think of a nice way to put my opencl patch in another file | 21:14 |
PhilTillet | :D | 21:15 |
@sonney2k | n4nd0, uh! I just got an email from JL - he is no longer at yahoo! | 21:20 |
n4nd0 | sonney2k: oh, is he already at another company? | 21:21 |
puffin444 | JL? | 21:21 |
n4nd0 | John Langford | 21:21 |
@sonney2k | puffin444, inventor of isomap | 21:22 |
@sonney2k | and author of the fastest covertree on earth | 21:22 |
n4nd0 | and the cover tree ;) | 21:22 |
n4nd0 | haha | 21:22 |
@sonney2k | and vowpal vabbit :) | 21:22 |
n4nd0 | we are still not done with ours :P | 21:22 |
@sonney2k | w | 21:22 |
gsomix | legend :) | 21:22 |
puffin444 | Oh isomap? The manifold learning algorithm? | 21:23 |
n4nd0 | puffin444: yeah, dimensionality reduction | 21:23 |
@sonney2k | puffin444, this guy http://hunch.net/~jl/ | 21:24 |
n4nd0 | I think he has also been with Thrun with particle filters | 21:24 |
puffin444 | Ok I knew that name sounded familiar | 21:24 |
puffin444 | Sebastian Thrun? | 21:24 |
n4nd0 | yeah | 21:24 |
puffin444 | Oh wow! | 21:24 |
n4nd0 | they are both authors of this paper http://hunch.net/~jl/projects/dt_particle_filter/nips02.pdf | 21:26 |
@sonney2k | he is a big shot and a cool guy (as you can guess from the picture :) | 21:26 |
n4nd0 | I think it is pretty awesome that he was mentor of a project in shogun last year | 21:27 |
n4nd0 | sonney2k: no project suited or not enough time this year? | 21:28 |
puffin444 | Oh wow he has also been involved in captcha | 21:28 |
@sonney2k | n4nd0, also I didn't have enough time this year | 21:30 |
puffin444 | Looking at his website, it looks like he has done a lot of cool stuff. It must have been exciting to have him as a mentor. | 21:30 |
@sonney2k | and am having | 21:30 |
@sonney2k | unfortunatly shashwat didn't continue development | 21:32 |
n4nd0 | sonney2k: what do you think about particle filters? could it be a place in shogun for it? | 21:32 |
@sonney2k | I would have hoped he does some more on vw / shogun | 21:32 |
n4nd0 | I'm just curious about it | 21:33 |
PhilTillet | wah, paper with Thrun! | 21:33 |
@sonney2k | I don't know anything about it | 21:33 |
@sonney2k | hmmhh I know JL better than Thrun (would even say he is more famous) | 21:33 |
n4nd0 | oh do you think so? | 21:33 |
@sonney2k | but I guess due to this stanford lecture thrun is more known | 21:33 |
n4nd0 | in the environment I mode Thrun is like a rock star :D | 21:33 |
@sonney2k | and of course for winning the darpa thing | 21:33 |
n4nd0 | I move* | 21:34 |
PhilTillet | yes :p that's how I knew him | 21:34 |
PhilTillet | that ai-class made him a rockstar lol | 21:34 |
n4nd0 | for the people in my department he has been a rockstar since fast slam | 21:34 |
@sonney2k | fast slam? | 21:34 |
n4nd0 | they have told me about a lecture he gave here at KTH about it | 21:34 |
@sonney2k | about? | 21:35 |
n4nd0 | about fast slam | 21:35 |
n4nd0 | it is a method used for localization in robotics | 21:35 |
n4nd0 | SLAM stands for simultaneous localization and mapping | 21:35 |
n4nd0 | a robot doesn't know where it is and goes around building a map and getting an idea where it is at the same time | 21:36 |
@sonney2k | ok | 21:37 |
n4nd0 | sonney2k: about particle filters, are you more familiar with the condensation algorithm? | 21:37 |
@sonney2k | ok then about rockstars | 21:37 |
@sonney2k | does anyone know gary bradski here? | 21:37 |
n4nd0 | yeah | 21:37 |
@sonney2k | n4nd0, not at all | 21:37 |
n4nd0 | Willow Garage, OpenCV | 21:38 |
puffin444 | Doen't ring a bell | 21:38 |
@sonney2k | yeah ok | 21:38 |
n4nd0 | sonney2k: I checked the videolecture with he and you ;) | 21:38 |
@sonney2k | leon bottou? | 21:38 |
n4nd0 | new for me | 21:38 |
@sonney2k | yann lecun? | 21:38 |
@sonney2k | hmmhh | 21:39 |
@sonney2k | alex smola? | 21:39 |
n4nd0 | yeah | 21:39 |
@sonney2k | bernhard schoelkopf? | 21:39 |
n4nd0 | haha | 21:39 |
@sonney2k | klaus-robert mueller :D | 21:39 |
n4nd0 | I don't know that many | 21:39 |
@sonney2k | n4nd0, ok then last one | 21:39 |
mohataher | I have a question, when to use SGVector and when to use SGMatrix ? (when having one column only of course), I ask coz I was advised in a pull-request to use SGMatrix instead of SGVector coz it's lightweight. Any explanation ? | 21:40 |
@sonney2k | vladimir vapnik? | 21:40 |
n4nd0 | sonney2k: SVMs!! | 21:40 |
PhilTillet | sonney2k, yes for Yann Lecun ! he's french after all :p | 21:40 |
@sonney2k | n4nd0, ok then you should know smola / schoelkopf too | 21:40 |
gsomix | http://www.lecun.org/gallery/libpro/20011121-allyourbayes/dsc01228-02-h.jpg | 21:40 |
@sonney2k | and actually klaus :) | 21:40 |
PhilTillet | lol gsomix | 21:41 |
@sonney2k | gsomix, hehe | 21:41 |
puffin444 | Yann Lecun sounds familiar | 21:41 |
n4nd0 | lately I see Joachims everywhere | 21:41 |
@sonney2k | Thorsten? | 21:41 |
n4nd0 | yes | 21:41 |
@sonney2k | n4nd0, where? | 21:41 |
n4nd0 | lot of papers and work related to SO | 21:42 |
@sonney2k | old ones then or? | 21:42 |
@sonney2k | JL goes to microsoft... | 21:42 |
n4nd0 | mmm I think 5 years old at most, are those considered old ones? | 21:43 |
@sonney2k | I guess they are in better shape than yahoo | 21:43 |
n4nd0 | oh microsoft | 21:43 |
@sonney2k | yes old | 21:44 |
n4nd0 | he is the author too of SVM struct | 21:44 |
-!- genix [~gsomix@188.168.5.212] has joined #shogun | 21:45 | |
@sonney2k | n4nd0, yeah that is all old stuff | 21:45 |
@sonney2k | he did svmlight in the 90's | 21:46 |
n4nd0 | SVM struct appears dated 2008 on the webpage | 21:46 |
@sonney2k | last update | 21:46 |
n4nd0 | I see | 21:46 |
-!- gsomix [~gsomix@188.168.14.33] has quit [Ping timeout: 252 seconds] | 21:47 | |
@sonney2k | I guess I am old known all these in person | 21:47 |
@sonney2k | bah | 21:47 |
@sonney2k | anyways they know me too :D | 21:47 |
@sonney2k | and now ... time to sleep | 21:48 |
@sonney2k | cu | 21:48 |
n4nd0 | sonney2k: you are also a rock star :) | 21:48 |
n4nd0 | good night | 21:48 |
-!- fman [b890a72d@gateway/web/freenode/ip.184.144.167.45] has joined #shogun | 21:48 | |
n4nd0 | sonney2k: btw, particle filters, are least nice to see in action | 21:48 |
n4nd0 | http://www.youtube.com/watch?v=nWvLX6xmoAw&feature=related | 21:48 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 21:53 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has joined #shogun | 22:04 | |
-!- wiking [~wiking@78-23-189-112.access.telenet.be] has quit [Changing host] | 22:04 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 22:04 | |
-!- puffin444 [230bf329@gateway/web/freenode/ip.35.11.243.41] has quit [Quit: Page closed] | 22:19 | |
-!- fman [b890a72d@gateway/web/freenode/ip.184.144.167.45] has left #shogun [] | 22:45 | |
-!- genix [~gsomix@188.168.5.212] has quit [Ping timeout: 245 seconds] | 23:31 | |
-!- genix [~gsomix@188.168.13.79] has joined #shogun | 23:33 | |
-!- genix is now known as gsomix | 23:33 | |
gsomix | wiking, yo | 23:33 |
-!- mohataher [3e8758c4@gateway/web/freenode/ip.62.135.88.196] has quit [Quit: Page closed] | 23:35 | |
--- Log closed Fri Apr 20 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!