--- Log opened Tue Apr 03 00:00:19 2012 | ||
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has joined #shogun | 00:02 | |
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has quit [Changing host] | 00:02 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 00:02 | |
-!- tibi_popa [tibi_popa@95.76.45.235] has left #shogun [] | 00:03 | |
n4nd0 | shogun-buildbot: oh, C# successful, good job! | 00:04 |
---|---|---|
shogun-buildbot | What you say! | 00:04 |
gsomix | shogun-buildbot, test. | 00:05 |
gsomix | :( | 00:05 |
gsomix | Ok. | 00:05 |
-!- harshit_ [~harshit@182.68.67.61] has quit [Ping timeout: 265 seconds] | 00:06 | |
-!- harshit_ [~harshit@182.68.67.61] has joined #shogun | 00:06 | |
n4nd0 | hehe | 00:07 |
n4nd0 | gsomix: how is it going with your proposal man | 00:07 |
n4nd0 | ? | 00:07 |
gsomix | n4nd0, i'm writing now. | 00:08 |
n4nd0 | gsomix: cool! do you find useful the one I sent you? | 00:08 |
gsomix | yep | 00:09 |
n4nd0 | :) | 00:09 |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has joined #shogun | 00:11 | |
gsomix | Oh. Jimi Hendrix and Duane Allman are ones of a greatest guitarists in history... I can listen to them a very long time. | 00:21 |
gsomix | But it's time to sleep. | 00:22 |
gsomix | Goog night. | 00:22 |
-!- karlnapf [~heiko@host86-181-9-49.range86-181.btcentralplus.com] has joined #shogun | 00:24 | |
harshit_ | n4nd0: please send me sample proposals also .. | 00:29 |
harshit_ | having sometrouble making mine :( | 00:29 |
n4nd0 | harshit_: trouble? why so man? | 00:29 |
harshit_ | For creating the detailed timeline | 00:30 |
harshit_ | what material were you talking about with gsomix ? | 00:30 |
n4nd0 | I can send you that part then | 00:30 |
n4nd0 | in any case, what projects are you applying for? | 00:30 |
n4nd0 | it would be kind of weird to share proposal for the same project :P | 00:30 |
harshit_ | C5.0 + regression solvers of liblinear | 00:30 |
harshit_ | I think yours is SO learning . right ? | 00:31 |
n4nd0 | ok, so give me an address and I can send you the timeline for the multiclass project I have done today ok? | 00:31 |
n4nd0 | or SO, the one you prefer, they are pretty similar in structure actually | 00:31 |
harshit_ | here : harshit.syal@gmail.com | 00:32 |
n4nd0 | but I warn you, I think that mine's not pretty detailed at all | 00:32 |
n4nd0 | it's quite basic actually :S | 00:32 |
harshit_ | but still i'll get an idea | 00:32 |
n4nd0 | yeah, sure | 00:33 |
n4nd0 | please use it just as an idea ... it would be weird if we all send a very similar proposal :P | 00:33 |
harshit_ | yeah sure , i wont copy it :) | 00:34 |
harshit_ | but i thought they ask for weekly timeline | 00:34 |
harshit_ | i mean details of each weak | 00:35 |
harshit_ | week* | 00:35 |
n4nd0 | I don't really know | 00:35 |
n4nd0 | I suggest you to ask S?ren to get a better idea what do they want | 00:35 |
n4nd0 | try to ask when you meet him around here so I can get also an answer for that :) | 00:36 |
harshit_ | sure ;) | 00:36 |
harshit_ | g2g its about 5 here , need to sleep | 00:37 |
n4nd0 | yeah sure | 00:37 |
n4nd0 | I was going to ask you about that man | 00:37 |
harshit_ | bye | 00:37 |
n4nd0 | becuase I see that you are with us in the chat lot of times | 00:37 |
n4nd0 | but you have another time ! | 00:37 |
harshit_ | yeah have to manage somehow.otherwise my work wont progress | 00:38 |
harshit_ | hasta luego | 00:38 |
-!- harshit_ [~harshit@182.68.67.61] has quit [Quit: Leaving] | 00:38 | |
-!- harshit_ [~harshit@182.68.67.61] has joined #shogun | 00:39 | |
harshit_ | forgot to mention, Thanks for sending me your timeline | 00:40 |
-!- harshit_ [~harshit@182.68.67.61] has quit [Client Quit] | 00:40 | |
n4nd0 | you are welcome, good night | 00:41 |
-!- av3ngr [av3ngr@nat/redhat/x-tfqcnnyvsaidfgdq] has joined #shogun | 00:50 | |
-!- karlnapf [~heiko@host86-181-9-49.range86-181.btcentralplus.com] has left #shogun [] | 01:03 | |
shogun-buildbot | build #425 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/425 | 01:19 |
n4nd0 | shogun-buildbot: great!! | 01:20 |
shogun-buildbot | What you say! | 01:20 |
-!- mohataher [52c9fc37@gateway/web/freenode/ip.82.201.252.55] has joined #shogun | 02:12 | |
mohataher | Hi, I'm mohamed, applying for GSoC and I have a question regarding the application form for it. Is there anyone to answer my question now ? :) | 02:13 |
-!- nickon [~noneedtok@dD5774105.access.telenet.be] has quit [Quit: ( www.nnscript.com :: NoNameScript 4.22 :: www.esnation.com )] | 02:19 | |
n4nd0 | mohataher: hey mohamed! maybe I can help you | 02:21 |
n4nd0 | tell me | 02:21 |
mohataher | Thank you. Regarding sending a pull request, does it have to include any modification to the code/documentation or just a to try it out ? | 02:57 |
n4nd0 | it should include a patch | 03:00 |
n4nd0 | I think there was a conversation about that in the mailing list | 03:01 |
n4nd0 | but definitely, an empty patch is not enough | 03:02 |
-!- LeStealth [~benjamin@CPE001921e5c431-CM00111ae68378.cpe.net.cable.rogers.com] has joined #shogun | 03:14 | |
mohataher | okay, thanks for your answer. I will search for it. | 03:15 |
n4nd0 | let me know if you need help to find it | 03:16 |
mohataher | i found the mailing list and I'm skimming it now. I will let you know if I need help. Thank you. | 03:21 |
mohataher | I found the answer. Thanks. | 03:24 |
-!- mohataher [52c9fc37@gateway/web/freenode/ip.82.201.252.55] has quit [Quit: Page closed] | 03:24 | |
-!- LeStealth [~benjamin@CPE001921e5c431-CM00111ae68378.cpe.net.cable.rogers.com] has left #shogun ["Leaving"] | 03:27 | |
-!- hpsoar_ [7c108901@gateway/web/freenode/ip.124.16.137.1] has joined #shogun | 03:37 | |
-!- hpsoar_ [7c108901@gateway/web/freenode/ip.124.16.137.1] has quit [Ping timeout: 245 seconds] | 03:44 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 03:45 | |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has quit [Quit: flxb] | 03:58 | |
-!- pythonroar [d2198538@gateway/web/freenode/ip.210.25.133.56] has joined #shogun | 04:32 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 05:07 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 05:26 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 05:52 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Remote host closed the connection] | 06:19 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 06:19 | |
-!- pythonroar [d2198538@gateway/web/freenode/ip.210.25.133.56] has quit [Ping timeout: 245 seconds] | 06:50 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 07:55 | |
gsomix | hi | 08:00 |
n4nd0 | hey | 08:02 |
@sonney2k | n4nd0, so is our buildbot happy again? | 08:07 |
n4nd0 | sonney2k: it looked like that tonight :) | 08:08 |
n4nd0 | let's see if it continues well | 08:08 |
n4nd0 | sonney2k: yay! everythin all right | 08:08 |
n4nd0 | everything* | 08:08 |
CIA-64 | shogun: Heiko Strathmann master * rbad4740 / src/shogun/features/Subset.cpp : small cast fix - http://git.io/HefGQQ | 08:10 |
CIA-64 | shogun: Soeren Sonnenburg master * r0e4f543 / src/shogun/features/Subset.cpp : | 08:10 |
CIA-64 | shogun: Merge pull request #413 from karlnapf/master | 08:10 |
CIA-64 | shogun: small SG_ADD type cast fix - http://git.io/_zoRkA | 08:10 |
@sonney2k | so shogun-buildbot work! | 08:11 |
@sonney2k | n4nd0, heh even warnings down to nothing :) | 08:11 |
n4nd0 | sonney2k: cool! | 08:11 |
sonne|work | n4nd0: btw how does the mkl picture look like? | 08:44 |
n4nd0 | sonne|work: this is for very few samples | 08:48 |
n4nd0 | sonne|work: http://dl.dropbox.com/u/11020840/mkl.png | 08:49 |
sonne|work | n4nd0: could you run this with shading interp ? | 08:49 |
n4nd0 | sonne|work: I think it is already set like that | 08:49 |
n4nd0 | sonne|work: http://snipt.org/ugjgd6, line 116 | 08:50 |
sonne|work | something is wrong then | 08:52 |
sonne|work | contours look pretty ok IMHO | 08:52 |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 08:53 | |
n4nd0 | aham | 08:53 |
sonne|work | I guess the first test to fix this would be to use just 1 kernel | 08:53 |
sonne|work | and if it still looks weird with 1 then figure out what the difference is... | 08:54 |
sonne|work | n4nd0: I just tried - looks still totally broken ... but no time to go deeper for now | 08:58 |
n4nd0 | sonne|work: what do you suggest to use a CombinedKernel just appending one more kernel or one PolyKernel alone? | 08:58 |
n4nd0 | sonne|work: yeah, I tried changing a couple of things and it broke too :S | 08:59 |
sonne|work | I even used a single GaussianKernel | 08:59 |
sonne|work | still all weird | 08:59 |
n4nd0 | aham | 08:59 |
n4nd0 | I can try to take it a look and show you the results in some minutes | 09:00 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 09:11 | |
-!- pythonroar [d2198538@gateway/web/freenode/ip.210.25.133.56] has joined #shogun | 09:19 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 09:37 | |
-!- blackburn [~qdrgsm@83.234.54.186] has joined #shogun | 09:42 | |
blackburn | uhuh | 09:44 |
n4nd0 | blackburn: trying to scare people? :P | 09:44 |
blackburn | no people scared me | 09:45 |
blackburn | moaaaar pull requests | 09:45 |
blackburn | hah | 09:45 |
n4nd0 | haha | 09:46 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 09:50 | |
-!- Marty28 [~Marty@158.181.76.57] has joined #shogun | 09:51 | |
Marty28 | hi | 09:51 |
n4nd0 | Marty28: hey! | 09:52 |
-!- harshit_ [~harshit@59.177.193.26] has joined #shogun | 09:57 | |
sonne|work | hey harshit_! | 10:10 |
sonne|work | nice work | 10:10 |
sonne|work | btw I didn't understand your benchmark numbers | 10:10 |
sonne|work | how long did it take on covertype? | 10:11 |
sonne|work | for shogun and how long with just matlab? | 10:11 |
-!- harshit_ [~harshit@59.177.193.26] has quit [Ping timeout: 246 seconds] | 10:14 | |
-!- harshit_ [~harshit@59.177.193.26] has joined #shogun | 10:19 | |
-!- blackburn [~qdrgsm@83.234.54.186] has quit [Ping timeout: 265 seconds] | 10:25 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 276 seconds] | 10:27 | |
-!- harshit_ [~harshit@59.177.193.26] has quit [Ping timeout: 244 seconds] | 10:32 | |
* gsomix went to classes about computer networks. | 10:34 | |
-!- harshit_ [~harshit@59.177.193.26] has joined #shogun | 10:49 | |
-!- harshit_ [~harshit@59.177.193.26] has quit [Quit: Leaving] | 11:03 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Remote host closed the connection] | 11:21 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 11:22 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Remote host closed the connection] | 11:22 | |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 11:23 | |
-!- karlnapf [~heiko@host86-181-9-49.range86-181.btcentralplus.com] has joined #shogun | 11:25 | |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has joined #shogun | 11:31 | |
-!- av3ngr [av3ngr@nat/redhat/x-tfqcnnyvsaidfgdq] has quit [Quit: That's all folks!] | 11:51 | |
-!- nickon [~noneedtok@dD5774105.access.telenet.be] has joined #shogun | 11:56 | |
-!- wiking [~wiking@iridia-pc21.ulb.ac.be] has joined #shogun | 12:09 | |
-!- wiking [~wiking@iridia-pc21.ulb.ac.be] has quit [Changing host] | 12:09 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 12:09 | |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has quit [Quit: flxb] | 12:26 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 12:26 | |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has joined #shogun | 12:28 | |
n4nd0 | I am taking a look to the seg fault Nicholas has pointed out | 12:48 |
-!- flxb [~cronor@e178173124.adsl.alicedsl.de] has quit [Quit: flxb] | 13:03 | |
n4nd0 | sonne|work: around? | 13:18 |
sonne|work | yes | 13:19 |
n4nd0 | the way to do CV for kernel machines is to use the constructor that takes a machine and no features right? | 13:20 |
n4nd0 | the kernel has to be set to that machine | 13:20 |
sonne|work | I think so - karlnapf is the expert for that | 13:24 |
karlnapf | hum? just read my name :) | 13:24 |
karlnapf | well you can specify a kernel before | 13:25 |
karlnapf | depends on what you want to do | 13:25 |
karlnapf | just simple CV for evaluation or parameter search? | 13:25 |
n4nd0 | karlnapf: I am trying to fix the example of the guy that sent the mail to the list | 13:26 |
n4nd0 | his problem is that when he does CV with LibSVM it seg faults | 13:26 |
karlnapf | ah, ok | 13:26 |
karlnapf | CV means what? | 13:26 |
n4nd0 | the kernel has never been set, we should lock that | 13:26 |
n4nd0 | cross validation, sorry | 13:26 |
karlnapf | just for simple evaluation or for a parameter search? | 13:27 |
n4nd0 | so now I am trying to find out how to do it for LibSVM, I have managed to get something running | 13:27 |
n4nd0 | I am afraid I cannot answer that :S | 13:27 |
n4nd0 | I can so you the code of the example if so and you can tell me if it is all right, ok? | 13:28 |
karlnapf | mmh, well | 13:28 |
karlnapf | yes sure | 13:28 |
n4nd0 | http://snipt.org/ugjie0 | 13:28 |
n4nd0 | so in the line 45 | 13:28 |
karlnapf | ok | 13:29 |
n4nd0 | before it was simply LibSVM(), and it turned out to seg fault for this issue that we get to execute a method for the kernel member of LibSVM | 13:29 |
karlnapf | yes you have to set the kernel before training in that case | 13:29 |
n4nd0 | and the kernel had never been initialized | 13:29 |
karlnapf | the parameter search does not involve the kernel at all | 13:29 |
n4nd0 | so this is cross validation then? | 13:29 |
karlnapf | no its a parameter search | 13:29 |
karlnapf | but this is based on cross-validation | 13:30 |
n4nd0 | aham | 13:30 |
karlnapf | since you have to test your current parameters somehow | 13:30 |
karlnapf | and thats done with CV | 13:30 |
n4nd0 | ok, I think I get the main idea | 13:30 |
karlnapf | there are two seperate classes for that, you can see them in the example | 13:30 |
n4nd0 | my only concern is that now to create the LibSVM | 13:30 |
karlnapf | mmh | 13:30 |
karlnapf | doesnt that example work? | 13:30 |
karlnapf | should | 13:30 |
n4nd0 | I have to use a value for C, and the labels | 13:31 |
n4nd0 | yes yes, it does | 13:31 |
n4nd0 | but idk if it does what it should do ;) | 13:31 |
n4nd0 | so the point is | 13:31 |
n4nd0 | when the instance of LibSVM is created now | 13:31 |
n4nd0 | this values, are used somewhen? | 13:31 |
karlnapf | which values? | 13:32 |
n4nd0 | the values for C and the labels | 13:32 |
karlnapf | ah ok | 13:32 |
karlnapf | well the labels are used in training, in the CV, they are needed | 13:32 |
n4nd0 | I mean it because C is supposed to be tried out within an interval | 13:32 |
karlnapf | and the C value is just overwritten | 13:32 |
n4nd0 | ok | 13:32 |
n4nd0 | but the labels are given again in the CV constructor, line 55 | 13:33 |
karlnapf | oh, yes, true | 13:33 |
karlnapf | well, then you could create the SVM with an empty constructor | 13:33 |
karlnapf | and set just the kernel by hand instead | 13:33 |
karlnapf | classifier=LibSVM() | 13:33 |
karlnapf | classifier.set_kernel(kernrel) | 13:33 |
n4nd0 | I think that is fairly reasonable | 13:33 |
karlnapf | (or so) | 13:33 |
karlnapf | we ineed do sometimes not check for good initialization | 13:34 |
n4nd0 | great | 13:34 |
karlnapf | thats currently all a bit complicated i guess :) | 13:34 |
n4nd0 | it works as well with the way you've suggested | 13:34 |
karlnapf | nice then | 13:35 |
n4nd0 | and I like it more than using the labels in the constructor | 13:35 |
n4nd0 | I found that confusing | 13:35 |
karlnapf | yes its nicer that way | 13:35 |
n4nd0 | ok | 13:35 |
karlnapf | say n4nd0, are you fernando? | 13:36 |
n4nd0 | so I think that I should just add a fix in the method that was seg faulting to fire SG_ERROR in case the kernel is null | 13:36 |
n4nd0 | karlnapf: yeah! you are Heiko right? | 13:36 |
karlnapf | I am a bit confused by all these nicknames | 13:36 |
karlnapf | yes :) | 13:36 |
n4nd0 | haha | 13:36 |
n4nd0 | but yours is even more difficult to guess :P | 13:37 |
karlnapf | yes, an error message if there is no kernelk would be nice | 13:37 |
karlnapf | yes, true ;) | 13:37 |
karlnapf | when blackburn shows up, we can talk about that subset stuff | 13:37 |
n4nd0 | karlnapf: sure | 13:37 |
n4nd0 | in the meantime I will add a pull request with the fix and answer Nicholas | 13:38 |
karlnapf | nice one | 13:38 |
karlnapf | I will continue on my work now, but will have one eye on the irc, see you! | 13:38 |
n4nd0 | thank you, bye! | 13:39 |
n4nd0 | sonne|work: jasmine's pull request looks nice! | 13:58 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 240 seconds] | 14:06 | |
-!- sidsniper [~piggy@14.139.82.6] has joined #shogun | 14:27 | |
CIA-64 | shogun: iglesias master * rce07314 / (2 files): * fix seg fault when trying to do CV with no kernel initialized, * fix indent - http://git.io/4es7wg | 14:30 |
CIA-64 | shogun: Soeren Sonnenburg master * r2b4bdc2 / (2 files): | 14:30 |
CIA-64 | shogun: Merge pull request #415 from iglesias/fix-modelsel | 14:30 |
CIA-64 | shogun: Small fix to avoid seg fault - http://git.io/vqs8lQ | 14:30 |
-!- sidsnipe [~piggy@14.139.82.6] has joined #shogun | 14:32 | |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 14:32 | |
-!- sidsniper [~piggy@14.139.82.6] has quit [Ping timeout: 260 seconds] | 14:33 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 14:33 | |
-!- sidsnipe [~piggy@14.139.82.6] has quit [Client Quit] | 14:33 | |
blackburn | karlnapf: hey there | 14:34 |
karlnapf | hey blackbutrn | 14:41 |
karlnapf | blackburn, hi | 14:41 |
blackburn | karlnapf: I heard you wanted to kill me because of machine subsetting"? ;) | 14:42 |
karlnapf | krkr ;) | 14:42 |
karlnapf | no, just asking about it | 14:42 |
karlnapf | explain it so me | 14:42 |
blackburn | ok | 14:42 |
blackburn | lets take a look | 14:42 |
blackburn | once you train machines OvO | 14:43 |
blackburn | you need to subset your trainset | 14:43 |
karlnapf | OvO is one versus one? | 14:43 |
sonne|work | yes | 14:43 |
blackburn | yes | 14:43 |
blackburn | https://github.com/shogun-toolbox/shogun/blob/master/src/shogun/machine/MulticlassMachine.cpp | 14:44 |
sonne|work | yes^2 | 14:44 |
blackburn | hah | 14:44 |
blackburn | 2yes | 14:44 |
karlnapf | hehe | 14:44 |
karlnapf | ok, why the need for subsets? | 14:44 |
blackburn | because it would be painful | 14:44 |
blackburn | to create new features | 14:44 |
blackburn | each time you want to construct ovo machine | 14:44 |
karlnapf | no I meant why needs OvO that? | 14:44 |
karlnapf | sorry doorbell | 14:44 |
blackburn | hmm because all the classes but some (m,n) pair should be ignored | 14:45 |
karlnapf | re | 14:46 |
blackburn | any better way would be welcome but I can't see any | 14:46 |
sonne|work | blackburn: so this is 'only' to avoid copying of data around right? | 14:47 |
blackburn | sonne|work: yes probably | 14:47 |
karlnapf | How does LibSVMMutlicalss do this? | 14:47 |
sonne|work | karlnapf: they have some kind of subset index internally | 14:47 |
karlnapf | ok | 14:47 |
karlnapf | and why not subsetting features? | 14:48 |
sonne|work | karlnapf: possible... | 14:48 |
sonne|work | BUT | 14:48 |
sonne|work | what about model selection then? | 14:48 |
karlnapf | blows up | 14:48 |
sonne|work | can we have subsets of subsets | 14:48 |
karlnapf | mmh | 14:48 |
sonne|work | yeah | 14:48 |
karlnapf | Yes, one solution would be to have multiple subsets | 14:49 |
sonne|work | if we had a function to say train(features, subset) it would work | 14:49 |
karlnapf | like we talked about one year ago or so | 14:49 |
karlnapf | yes | 14:49 |
karlnapf | thats possible on the locked data | 14:49 |
sonne|work | or a function that returns a subset given a subset as input | 14:49 |
karlnapf | but thats not good for OvO | 14:49 |
sonne|work | that might be sufficient ... | 14:49 |
karlnapf | I mean | 14:49 |
sonne|work | so a function that computes a new subset based on the current subset | 14:49 |
karlnapf | modelselection on kernel machines without locking is a pain anyway | 14:49 |
sonne|work | only if you can store the kernel in memory... | 14:50 |
sonne|work | (that was never the case in the applications I had) | 14:50 |
karlnapf | yes true | 14:50 |
karlnapf | ok so subsetting the machine | 14:51 |
karlnapf | that seems to be a good solution, however, we got many subset mechanisms then | 14:51 |
karlnapf | and they basically all take subsets of features | 14:51 |
karlnapf | if there was a way to register multiple subsets on features | 14:51 |
karlnapf | we wouldnt need all that stuff | 14:51 |
sonne|work | ok true | 14:52 |
karlnapf | like give each subset a unique identifier | 14:52 |
karlnapf | and pass it before accessing them | 14:52 |
sonne|work | some dynarray of subsets... | 14:52 |
karlnapf | yes | 14:52 |
sonne|work | but actually that would kill performance | 14:52 |
karlnapf | and how to access? | 14:52 |
sonne|work | maybe one can add subsets as one wants | 14:53 |
karlnapf | features subsets are currently used implicitly | 14:53 |
sonne|work | and then a 'current' subset is computed based on the subsets in the list | 14:53 |
sonne|work | so accessing features uses only this 'current' subset | 14:53 |
sonne|work | maybe current is the wrong term 'merged subset' is better | 14:53 |
karlnapf | oh yes, they would have a hierachy | 14:54 |
karlnapf | hierarchy | 14:54 |
karlnapf | still the problem: | 14:54 |
karlnapf | A sets subset A | 14:54 |
karlnapf | B sets another subset on that | 14:54 |
karlnapf | B is not ready yet | 14:54 |
karlnapf | A wants to access | 14:54 |
karlnapf | and because A is just setting the subset once and then accesses implicitly this is a problem | 14:55 |
sonne|work | well I would say that B has to free its subset before passing on to A | 14:55 |
sonne|work | otherwise this gets too difficult | 14:55 |
karlnapf | ok | 14:55 |
blackburn | I do not really understand why you need any hierarchy or other stuff | 14:56 |
karlnapf | then it would be quite straightforward to implement right? | 14:56 |
blackburn | what are you trying to solve right now? | 14:56 |
karlnapf | blackburn, the problem is that modelselection subsets features | 14:56 |
karlnapf | and if the multiclass svm also does that, there is a kind of conflict | 14:56 |
karlnapf | a "subset on a subset" | 14:56 |
blackburn | oh | 14:56 |
karlnapf | would be needed, but currently thats not possible | 14:57 |
blackburn | yes got it | 14:57 |
karlnapf | and having multiple subset mechanisms (features, multiclass, locking) is a bit ugly | 14:57 |
sonne|work | karlnapf: yes I think this is doable | 14:57 |
blackburn | multiclass do not do subsetting actually | 14:57 |
blackburn | it subsets features | 14:57 |
blackburn | or kernel | 14:57 |
blackburn | nothing moer | 14:57 |
sonne|work | just some 'add_subset' / 'remove_subset' functions | 14:57 |
sonne|work | and whenever add/remove is called the 'merged' subset is updated (which is what we currently have) | 14:58 |
sonne|work | so no need to change the code we have now - just some additions | 14:58 |
blackburn | subset union subset? | 14:58 |
sonne|work | actually one needs to change all code that sets a subset | 14:58 |
sonne|work | it needs to use add/remove_subset | 14:58 |
blackburn | each time you add subset it subsets the subset? | 14:58 |
sonne|work | blackburn: it computes the subset of the subset of the subset - yes | 14:59 |
blackburn | sonne|work: it may break things then | 14:59 |
-!- harshit_ [~harshit@182.68.249.161] has joined #shogun | 14:59 | |
karlnapf | we should set a maximum subset level to avoid abuse of this system ;) | 14:59 |
blackburn | earlier it was assumed that there is only one subset | 14:59 |
sonne|work | blackburn: what does it break? | 15:00 |
blackburn | sonne|work: no idea but karlnapf could know probably | 15:00 |
karlnapf | you lost me, what? | 15:00 |
sonne|work | karlnapf: blackburn is asking when setting subsets of subsets would break in the current system | 15:00 |
karlnapf | no I think not | 15:01 |
blackburn | ok then | 15:01 |
karlnapf | one just HAS to remove subsets after doing work on them | 15:01 |
blackburn | one issue is that add and remove are not interchangeable or so | 15:01 |
sonne|work | yeah so it should be safe | 15:01 |
blackburn | I mean add applies it on top of that | 15:01 |
blackburn | but remove removes all of them | 15:01 |
sonne|work | no | 15:02 |
blackburn | or I got wrong? | 15:02 |
sonne|work | remove removes 1 | 15:02 |
blackburn | you want to do stack there? | 15:02 |
sonne|work | yes | 15:02 |
karlnapf | yes | 15:02 |
harshit_ | sonne|work : I have updated the pull request, I think now its ready to be merged | 15:02 |
blackburn | good idea probably | 15:02 |
sonne|work | just an array of subsets | 15:02 |
karlnapf | Perhaps this removing could be possibly done automatically in some destructor | 15:02 |
blackburn | I thought you wanted to merge subsets actually | 15:02 |
sonne|work | harshit_: Thanks - I will have a look later | 15:02 |
sonne|work | blackburn: yes for performance | 15:02 |
karlnapf | then we would not have seperate subset systems at three places | 15:03 |
sonne|work | we have one 'merged' subset | 15:03 |
blackburn | well two subsets is ok probably | 15:03 |
sonne|work | which is computed from the array of subsets | 15:03 |
sonne|work | such that we don't have to go through the whole array of subsets all the time | 15:03 |
blackburn | karlnapf: which three places? | 15:03 |
karlnapf | yes, all the subset instances are saved but everytime the stack is touched, atemporary "current" one is created for performance thight? | 15:03 |
karlnapf | modsel, locking, multiclass | 15:03 |
blackburn | well I see no other way for multiclass ovo | 15:04 |
karlnapf | But when we have this new thing, you can just subset the features before training the class pairs | 15:05 |
karlnapf | and then the machines just train on "all" data (from their perspective) | 15:05 |
sonne|work | fight karlnapf fight | 15:05 |
sonne|work | :) | 15:05 |
karlnapf | lol ;) | 15:05 |
karlnapf | I went kick-boxing yesterday, still have Muskelkater, so, no thanks ;) | 15:06 |
blackburn | karlnapf: actually I think subset stack is a good idea not only wrt to ovo | 15:06 |
sonne|work | I also like it | 15:06 |
sonne|work | so who wants to implement it? | 15:06 |
blackburn | i.e. what if you want to do cross-validation on subset of your features? | 15:06 |
karlnapf | thats possible then | 15:07 |
sonne|work | blackburn: features -> you mean subset of dimensions | 15:07 |
sonne|work | or subset of examples? | 15:07 |
blackburn | oh no sorry | 15:07 |
blackburn | feature vectors | 15:07 |
karlnapf | that would work then | 15:07 |
sonne|work | yes | 15:07 |
blackburn | subsets of features would be interesting however it is a different system | 15:07 |
sonne|work | that's fine | 15:07 |
karlnapf | you just would have to check for class indices everytime, not just once in the beginning | 15:07 |
blackburn | is it required btw? | 15:08 |
karlnapf | yes, subsets of dimensions would be cool for feature selection | 15:08 |
karlnapf | but thats probably another story | 15:08 |
karlnapf | I cou?d implement the stuff | 15:08 |
blackburn | sonne|work: btw what do you think about sparse linear machine? | 15:08 |
blackburn | karlnapf: please do if you have time ;) | 15:09 |
blackburn | I can but you would probably be able to test it better | 15:09 |
karlnapf | Its not ultra urgent, is it? | 15:09 |
sonne|work | blackburn: what for? what does it do? | 15:09 |
blackburn | karlnapf: no, not urgent at all | 15:09 |
blackburn | sonne|work: lasso | 15:09 |
sonne|work | blackburn: please elaborate | 15:10 |
karlnapf | because my parents are visiting me tomorrow and I only have a few hours per day, and also want to come up with the hypothesis test interfaces this week | 15:10 |
blackburn | sonne|work: lasso has only a few non zeroes in [w] | 15:10 |
sonne|work | karlnapf: but it should be rather easy | 15:10 |
sonne|work | maybe n4ndo would want to do it? | 15:10 |
blackburn | or gsomix? | 15:10 |
sonne|work | I mean it would just mean to add these to add/remove functions | 15:10 |
sonne|work | blackburn: no gsomix is busy with python3 | 15:11 |
karlnapf | its ok, Ill do it | 15:11 |
sonne|work | and the merge function | 15:11 |
sonne|work | that's all or? | 15:11 |
karlnapf | there might be trouble with x-val | 15:11 |
blackburn | sonne|work: do you have any other tasks for gsomix? | 15:11 |
sonne|work | ...can be fixed when it occurs :) | 15:11 |
-!- harshit_ [~harshit@182.68.249.161] has quit [Quit: Leaving] | 15:11 | |
sonne|work | blackburn: e.g. octave_modular typemap for sparse features | 15:11 |
blackburn | sonne|work: because I want somebody to do LARS and he would do that | 15:11 |
blackburn | I see - then let him do this octave | 15:12 |
sonne|work | blackburn: you still didn't explain what sparselinearmachine is! | 15:12 |
sonne|work | gsomix: are you busy? | 15:12 |
blackburn | sonne|work: yes,yes - what if w is sparse? | 15:12 |
karlnapf | sonne|work, I read the mail yesterday about the parameter selection for GPs, I think there might be some potential for creating new model selection classes | 15:13 |
sonne|work | blackburn: ahh you mean not to store whole w vector | 15:13 |
blackburn | rather not to compute outputs with whole one | 15:13 |
blackburn | it is the case of lasso | 15:13 |
sonne|work | blackburn: or one just adds a sparsify_w function ... | 15:13 |
karlnapf | one could generalize the modsel class to use specific evaluation functions (like model evidence or LOO) and then also specify gradients of them and create new GradientDescentModelSelection class or so | 15:14 |
sonne|work | for training it is tough... | 15:14 |
blackburn | sonne|work: sparse training you mean? | 15:14 |
blackburn | yes no need to | 15:14 |
sonne|work | karlnapf: you would probably have to help this student then | 15:14 |
karlnapf | yes, I like that kind of stuff | 15:15 |
sonne|work | blackburn: one way would be to then call compress_w and then apply uses the compressed representation of w | 15:15 |
blackburn | SNOW | 15:15 |
blackburn | WTF | 15:15 |
blackburn | SNOW in april! | 15:15 |
karlnapf | hehe :) | 15:15 |
sonne|work | we just had that on sunday | 15:15 |
blackburn | it is all dirty here | 15:15 |
karlnapf | I got foggy rainy Londonish sky here :) | 15:15 |
blackburn | and snow again | 15:16 |
blackburn | what the fog! | 15:16 |
karlnapf | there was snow in germany? | 15:16 |
blackburn | sonne|work: ok then lars goes to n4ndo probably | 15:17 |
blackburn | I hope I do not overload him :D | 15:17 |
sonne|work | blackburn: if you know better what gsomix is interested in - tell me - or gsomix tell me :) | 15:18 |
blackburn | sonne|work: he is at uni right now probably | 15:18 |
blackburn | I saw him hour ago | 15:19 |
blackburn | sonne|work: director stuff btw | 15:20 |
blackburn | sonne|work: what is # of proposals? | 15:20 |
sonne|work | blackburn: he is more fond of that? | 15:20 |
sonne|work | directors I mean | 15:21 |
blackburn | sonne|work: I do not know | 15:21 |
blackburn | I'll ask | 15:21 |
blackburn | or you | 15:21 |
sonne|work | it is tough though and I guess he would need to write an isolated thing | 15:21 |
sonne|work | blackburn: 21 proposals by now | 15:22 |
sonne|work | 3 days left... | 15:22 |
blackburn | I think new candidates should finish some small successful tasks | 15:22 |
sonne|work | yes - even after deadline but before we decide | 15:22 |
blackburn | that's why I am afraid of sparse matrix and other stuff | 15:23 |
blackburn | can you estimate men-day's? :D | 15:23 |
blackburn | man* | 15:24 |
sonne|work | sparse matrix? | 15:24 |
sonne|work | what? | 15:24 |
blackburn | sonne|work: software eng managers used to estimate man hours and man days :D | 15:25 |
blackburn | sonne|work: sparse matrix support for octave | 15:25 |
sonne|work | that is very easy (for me) | 15:26 |
sonne|work | so lets say 2 days max | 15:26 |
blackburn | ok nice | 15:27 |
blackburn | sonne|work: will you have time to guide him? | 15:27 |
karlnapf | hey guys, Do you know this possibility of approximating a kernel matric using incomplete cholesky factorisation? | 15:27 |
karlnapf | matrix | 15:28 |
karlnapf | I recently used that and found it quite cool, so perhaps there would be some room for it in shogun | 15:28 |
blackburn | karlnapf: what is *incomplete* cholesky? | 15:28 |
blackburn | ah let me google that | 15:28 |
karlnapf | sent you an email | 15:29 |
karlnapf | its an extract from the kernel methods book of john shawe taylor and nello crist.... | 15:30 |
blackburn | :) | 15:30 |
blackburn | oh I'll check it later | 15:31 |
blackburn | karlnapf: did you send the proposal alread | 15:31 |
blackburn | y? | 15:31 |
karlnapf | yes | 15:32 |
karlnapf | hope you like it :) | 15:32 |
blackburn | karlnapf: this year we are going to receive 20% of last year ones :D | 15:32 |
karlnapf | what do you mean? | 15:32 |
karlnapf | ah like re-applying? :) | 15:33 |
blackburn | no | 15:34 |
blackburn | in total count | 15:34 |
blackburn | karlnapf: only you and me are re-applying afaik | 15:34 |
sonne|work | last year we had like 70 applications - this year it will be around 30 I guess | 15:35 |
karlnapf | really? how does that come? | 15:35 |
sonne|work | this is because we required patches even before applying | 15:35 |
karlnapf | projects too hard? | 15:35 |
sonne|work | and applications/candidates are much stronger | 15:35 |
karlnapf | mmh, was that a good idea then? the number of slots is determined by the number of applications thight? | 15:35 |
blackburn | not only I think | 15:36 |
sonne|work | yeah from that perspective not - but who can wade through 70+ applications? | 15:36 |
sonne|work | even reading them is *a lot* of work | 15:36 |
blackburn | sonne|work: especially if they are done by russians | 15:36 |
karlnapf | indeed | 15:36 |
blackburn | :D | 15:36 |
karlnapf | hehe :) | 15:37 |
sonne|work | and last year we had like 3 reviewers per student app | 15:37 |
sonne|work | and scores / final ranking | 15:37 |
blackburn | sonne|work: who did the job? | 15:37 |
sonne|work | we split work somehow | 15:37 |
blackburn | sonne|work: I mean who have participated in that? | 15:38 |
sonne|work | candidates that we were more sure of we wanted were reviewed by more people then | 15:38 |
sonne|work | everyone | 15:38 |
blackburn | mentors? | 15:38 |
sonne|work | of course | 15:38 |
sonne|work | mentors | 15:38 |
sonne|work | only | 15:38 |
sonne|work | that is how it works | 15:38 |
blackburn | I know, I am just curious :) | 15:38 |
sonne|work | I wish blackburn / karlnapf you both would be mentors instead of students | 15:38 |
sonne|work | :D | 15:38 |
karlnapf | Hopefully next year :) | 15:39 |
blackburn | sonne|work: if it was same-rewarded | 15:39 |
karlnapf | I desperately need money, unfortunately | 15:39 |
karlnapf | so expensive here | 15:39 |
sonne|work | no reward unfortuantely | 15:39 |
blackburn | yes that's why we want to be students :) | 15:40 |
* sonne|work wants to be student to | 15:40 | |
sonne|work | too | 15:40 |
-!- pythonroar [d2198538@gateway/web/freenode/ip.210.25.133.56] has quit [Ping timeout: 245 seconds] | 15:40 | |
karlnapf | How is work at tomtom btw? | 15:40 |
sonne|work | they have great data... | 15:41 |
sonne|work | and lots of | 15:41 |
sonne|work | and even more data analysis problems | 15:41 |
karlnapf | what are you actually doing there? Like what kind of problems? | 15:41 |
sonne|work | so quite interesting | 15:41 |
blackburn | sonne|work: I bet your year salary is larger than gsoc reward ;) | 15:41 |
sonne|work | I am working in the HD traffic team - that is the one predicting jams / any kind of live traffic stuff | 15:42 |
blackburn | sonne|work: btw can I intern there? ;) | 15:43 |
sonne|work | blackburn: you should go to university and become professor | 15:43 |
blackburn | would be funny | 15:44 |
blackburn | sonne|work: why should I? | 15:44 |
sonne|work | otherwise we will never have students to work for you on shogun | 15:44 |
karlnapf | lol :) | 15:44 |
blackburn | hah good reason to devote whole life for useless stuff there | 15:44 |
sonne|work | karlnapf: nothing to laugh here ;) you have the same destiny | 15:46 |
blackburn | sonne|work: what about you then? | 15:46 |
sonne|work | I left academia... | 15:46 |
karlnapf | we are all converging on growing shogun students ;) | 15:47 |
blackburn | yes I heard that before but no way back? | 15:47 |
blackburn | actually we've got one kind of brilliant haha | 15:47 |
blackburn | sonne|work: what I think is there is no need to worry we are not the mentors | 15:49 |
sonne|work | karlnapf: I am attempting to grow+educate 2 kids to become shogun developers :D | 15:49 |
blackburn | first shogun girl? | 15:49 |
blackburn | we would have to wait at least 15 years more though | 15:50 |
sonne|work | blackburn: I do - every help is welcome (as always) | 15:50 |
blackburn | sonne|work: you do worry you mean? | 15:50 |
sonne|work | nowadays programmers start 6 | 15:50 |
sonne|work | yes | 15:50 |
karlnapf | hehe, your kids | 15:50 |
karlnapf | I got a BASIC interpreter for my AMIGA for my 10th birthday from my father ;) | 15:51 |
blackburn | sonne|work: I think being mentors would mean we have no time for develop.. | 15:51 |
blackburn | I got my first computer when I was 7 | 15:51 |
karlnapf | And my parents always wanted me to do computer science, but I refused and went studying jazz guitar first ;) so dont put too much pressure up, sonne|work ;) | 15:53 |
sonne|work | karlnapf: my son is 11 months and want to play games on my android phone :) | 15:55 |
sonne|work | but I don't understand how I got interested in this | 15:55 |
blackburn | sonne|work: he is also used to multiclass stuff already | 15:55 |
sonne|work | maybe because my parents always wanted to limit computer-time | 15:56 |
karlnapf | hehe, dont let him become a super-nerd ;) | 15:56 |
karlnapf | yes same here | 15:56 |
sonne|work | so I guess that is the key | 15:56 |
sonne|work | tough limits (he wants to break) | 15:56 |
blackburn | sonne|work: you should deny him to code perfectly later | 15:57 |
* sonne|work thinks that karlnapf would have finished this sub-subsetting business by now if we wouldn't fool around... | 15:57 | |
sonne|work | hehe | 15:57 |
karlnapf | I am studying ;) | 15:58 |
blackburn | and yes deny him to construct theory better than vapnik did | 15:58 |
karlnapf | but youre right, I should stop chatting | 15:58 |
karlnapf | take care guys :) (wave) | 15:58 |
blackburn | my son, never develop general pattern recogntion theory better than VC and PAC | 15:58 |
blackburn | sonne|work: that's the thing you should say everyday ;) | 15:59 |
sonne|work | karlnapf: hope to see you in chat more often (idling) | 15:59 |
sonne|work | cu | 15:59 |
sonne|work | yeah | 16:00 |
karlnapf | yes, will leave it on more, uni just stopped so I can be online all day : bye | 16:00 |
blackburn | snow stopped | 16:03 |
blackburn | yay! | 16:04 |
sonne|work | karlnapf: bye | 16:04 |
blackburn | sonne|work: have you used clearcase ever? | 16:05 |
sonne|work | no | 16:05 |
blackburn | if someone asks you to use it | 16:05 |
blackburn | better run | 16:05 |
blackburn | or hide somewhere | 16:05 |
blackburn | I heard it is legal to kill person that suggests to use clearcase | 16:06 |
-!- pluskid [chatzilla@111.120.76.142] has joined #shogun | 16:08 | |
-!- flxb_ [~cronor@141.23.80.206] has joined #shogun | 16:09 | |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has quit [Ping timeout: 264 seconds] | 16:11 | |
-!- flxb_ is now known as flxb | 16:11 | |
blackburn | sonne|work: that paper I showed to you is to appear in shity 'computer optics' 36(2) :D | 16:15 |
sonne|work | computer optics??? | 16:16 |
blackburn | yes that is the name of journal.. | 16:17 |
blackburn | sonne|work: do you think I am crazy? I am not welcome to Russia :D | 16:17 |
blackburn | am not,* | 16:18 |
sonne|work | the world is crazy... | 16:18 |
blackburn | sonne|work: it is a kind of local journal with rather easy submission | 16:20 |
blackburn | sonne|work: and I needed it to 'power up' my bachelor's work | 16:20 |
blackburn | the most funny thing there is abstracts | 16:21 |
blackburn | 'In this paper is considered a method of forming a sequence of reliefs based on the specified reference terrain images by using a criterion based on computation a conjugation indexes between a current reliefs and a reference reliefs. An example of the implementation is given.' | 16:21 |
sonne|work | some lower machine learning conference would also have worked | 16:21 |
sonne|work | anyways got work to do cu | 16:21 |
blackburn | sonne|work: can you believe it is *real* abstract? | 16:21 |
blackburn | oh sure | 16:21 |
blackburn | see you | 16:21 |
blackburn | conference would work but no time for that | 16:21 |
wiking | blackburn i've decoded the email!!! :DDD | 16:23 |
blackburn | wiking: hahah | 16:23 |
blackburn | took a while right? | 16:24 |
wiking | well yeah | 16:24 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 16:24 | |
blackburn | n4nd0: can you remind me what is your 'contribution roadmap'? ;) | 16:25 |
blackburn | wiking: so what does he want? | 16:25 |
wiking | a little bit of change | 16:25 |
wiking | not so much | 16:25 |
wiking | basically we would remove the whole CLatentFeatures | 16:25 |
n4nd0 | blackburn: what do you mean, contributions in shogun? | 16:25 |
blackburn | n4nd0: yes | 16:25 |
wiking | and just have CFeatures | 16:25 |
wiking | but they would contain all the features: x,y,h | 16:26 |
blackburn | wiking: and include things to latentlabels? | 16:26 |
blackburn | ah | 16:26 |
blackburn | how to separate it when? | 16:26 |
wiking | well you don't need to | 16:26 |
blackburn | then* | 16:26 |
n4nd0 | blackburn: from the very beginning, CPLEX compilation, Mahalanobis distance, QDA, OvO, SPE | 16:26 |
wiking | optimizer doesn't need to know it | 16:26 |
n4nd0 | blackburn: not that much :S | 16:26 |
wiking | what part of the feature is what | 16:26 |
wiking | the only thing that after the results | 16:26 |
wiking | the user of the tool will have to know what is what | 16:27 |
blackburn | n4nd0: sorry I meant what is next? | 16:27 |
n4nd0 | blackburn: aham! | 16:27 |
blackburn | wiking: ah I see | 16:28 |
n4nd0 | blackburn: so first we should manage to get merged SPE, did you see I updated it? | 16:28 |
wiking | and then still one _could_ define an own function to speed up the optimization algo | 16:28 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Ping timeout: 276 seconds] | 16:28 | |
blackburn | wiking: that is hardly implied from the msg :D | 16:28 |
n4nd0 | blackburn: later, what about the OvO issue? I see you have been discussing long about it | 16:28 |
wiking | but if it's not provided it should just do a brute force search | 16:28 |
wiking | i'm just wondering now | 16:28 |
blackburn | n4nd0: we shall wait for new subset stack concept ready | 16:29 |
wiking | whether the whole class should be inherited.... i.e. have a protected virtual function for finding the argmax of the latent variable | 16:29 |
n4nd0 | blackburn: ok | 16:29 |
blackburn | n4nd0: I've got a new task for you thought, LARS | 16:29 |
blackburn | interested?;) | 16:29 |
blackburn | wiking: whole class of? | 16:29 |
-!- flxb [~cronor@141.23.80.206] has quit [Remote host closed the connection] | 16:29 | |
wiking | or have a class that only has a function like this.. and that class should be inherited and redefine that function for finding the argmax | 16:29 |
wiking | blackburn: LatentLinearMachine | 16:29 |
blackburn | aha | 16:30 |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has joined #shogun | 16:30 | |
n4nd0 | blackburn: tell me about it | 16:30 |
wiking | any input on it? | 16:30 |
wiking | i'd rather go with the latter option | 16:30 |
n4nd0 | blackburn: http://www.machinelearning.org/proceedings/icml2005/papers/053_GeneralizedLARS_Keerthi.pdf?? | 16:30 |
wiking | it feels better for me | 16:30 |
blackburn | n4nd0: generalized? simple would work as well :) | 16:30 |
blackburn | wiking: then go with latter ;) | 16:31 |
n4nd0 | blackburn: give me a pointer if so | 16:31 |
blackburn | n4nd0: scikits learn has one | 16:31 |
blackburn | n4nd0: I'll take a look on spe later | 16:31 |
n4nd0 | blackburn: cool | 16:31 |
pluskid | n4nd0: maybe this one? http://www.stanford.edu/~hastie/Papers/LARS/LeastAngle_2002.pdf | 16:32 |
blackburn | n4nd0: https://github.com/scikit-learn/scikit-learn/blob/master/sklearn/linear_model/least_angle.py | 16:32 |
blackburn | yes this one | 16:32 |
pluskid | you wish to include LARS in shogun for, solving LASSO ? | 16:32 |
blackburn | yes | 16:32 |
n4nd0 | regression, cool | 16:32 |
pluskid | LARS is cool, but I heard that there are many more efficient algorithm for L1 regularized problems? | 16:33 |
n4nd0 | aham, e.g.? | 16:33 |
blackburn | pluskid: I have not much expertise with linear regression, what would you suggest? | 16:33 |
pluskid | I do not know the literature (not familiar with optimization), actually, I always use LARS in my own work :p | 16:34 |
pluskid | blackburn: I heard some discussion about something like ??? gradient descent algorithms, but I do not understand the details | 16:35 |
n4nd0 | I know gradient descent | 16:35 |
pluskid | But I do think LARS has an advantage that it gives you the whole path | 16:35 |
pluskid | I think many other algorithms only gives you the solution for a particular \lambda (regularization coefficient) | 16:35 |
pluskid | Maybe any of you could check out this package: http://www.public.asu.edu/~jye02/Software/SLEP/ , if you are good at optimization | 16:36 |
blackburn | wow | 16:37 |
blackburn | nice reference | 16:37 |
blackburn | hmm that's something I should take care of | 16:38 |
blackburn | group based lasso is of interest for my project | 16:38 |
pluskid | hmm, I think those ideas interesting, too. But I never used them in practice yet, except simply L1 | 16:39 |
pluskid | simple | 16:39 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 16:40 | |
pluskid | Trying to find the discussion thread about LASSO optimization I have read before | 16:41 |
pluskid | Got it! | 16:43 |
pluskid | I list some references here, in case any one would be interested in. | 16:44 |
pluskid | They said that this algorithm is very fast: http://www-stat.stanford.edu/~jhf/ftp/glmnet.pdf | 16:44 |
pluskid | And for really huge data, online learning algorithms also exist: | 16:44 |
pluskid | http://jmlr.csail.mit.edu/papers/v10/langford09a.html | 16:44 |
pluskid | http://jmlr.csail.mit.edu/papers/v10/duchi09a.html | 16:44 |
pluskid | That's it. :) | 16:46 |
blackburn | pluskid: thanks! nice references | 16:46 |
pluskid | Time for sleep here. | 16:46 |
pluskid | good night/afternoon/morning guys! | 16:46 |
-!- pluskid [chatzilla@111.120.76.142] has quit [Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120312181643]] | 16:47 | |
blackburn | oh | 16:47 |
blackburn | good night | 16:47 |
n4nd0 | good sth ;) | 16:49 |
n4nd0 | so after the ideas that you suggested yesterday and the new one today, LARS ... | 16:51 |
n4nd0 | let's go for LARS because it is the newest :) | 16:51 |
blackburn | n4nd0: what have I suggested yesterday? | 16:52 |
n4nd0 | the tree | 16:53 |
n4nd0 | the cover tree for knn too | 16:53 |
blackburn | aha | 16:53 |
n4nd0 | parzen windows | 16:54 |
n4nd0 | ... | 16:54 |
n4nd0 | :P | 16:54 |
blackburn | n4nd0: what about parzen window one with covertree internally? later we would merge good design from it to knn | 16:55 |
n4nd0 | blackburn: idk if you're serious now | 16:55 |
blackburn | yes I am | 16:56 |
n4nd0 | so parzen windows can be implemented with cover trees?? | 16:56 |
blackburn | good question | 16:56 |
blackburn | probably not | 16:58 |
blackburn | :D | 16:58 |
blackburn | n4nd0: actually it could if there was some threshold based covertree | 17:00 |
blackburn | but it would require patching covertree that is present in shogun | 17:00 |
n4nd0 | blackburn: let's do that new threshold based covertree and write a paper :P | 17:02 |
blackburn | I am afraid it is way too simple idea | 17:02 |
blackburn | original covertree supports it IIRC | 17:02 |
@sonney2k | blackburn, is covertree only for euclidian distances? | 17:05 |
@sonney2k | or can it work with any distance? | 17:05 |
blackburn | sonney2k: any | 17:05 |
blackburn | it requires nothing but pairwise distances | 17:05 |
@sonney2k | ok then n4nd0 this definitely is my favourite then covertree knn :) | 17:06 |
n4nd0 | i'll focus on that then sonney2k | 17:08 |
CIA-64 | shogun: Harshit Syal master * r85ffdcd / src/shogun/classifier/svm/NewtonSVM.cpp : removed unwanted SG_SPRINT - http://git.io/_YLBPA | 17:09 |
CIA-64 | shogun: Soeren Sonnenburg master * ra934dcb / (7 files in 5 dirs): | 17:10 |
CIA-64 | shogun: Merge pull request #409 from harshitsyal/master | 17:10 |
CIA-64 | shogun: Newton SVM - http://git.io/EAZk4Q | 17:10 |
n4nd0 | I'll be back later, see you guys | 17:15 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 17:15 | |
@sonney2k | shogun-buildbot, ? | 17:15 |
-!- sidsniper [~piggy@14.139.82.6] has joined #shogun | 17:17 | |
sidsniper | hi blackburn | 17:19 |
blackburn | hi | 17:19 |
sidsniper | how can i test dim red algos? is there any data? | 17:19 |
blackburn | sidsniper: what kind of testing you want to do? | 17:19 |
sidsniper | reducing dimension of feature vector | 17:20 |
sidsniper | how did you test the correctness of algoriths implemented? | 17:20 |
blackburn | manually ;) and compared to other implementations | 17:21 |
sidsniper | so there must be some dataset which u had tested them upon? | 17:21 |
blackburn | yes, I tested it on swissroll, ORL and MIT-CBCL | 17:22 |
blackburn | swissroll is the simplest test I'd say | 17:22 |
sidsniper | okay | 17:23 |
sidsniper | so these are availaible online right? | 17:23 |
blackburn | sidsniper: yes | 17:23 |
sidsniper | ok thanks :) | 17:23 |
blackburn | sidsniper: have you doubts about its correctness? | 17:23 |
sidsniper | i have yet to test them..i m sure they will be correct :) | 17:24 |
blackburn | I'm going to implement unit-tests as well.. | 17:24 |
sidsniper | if you can give me links from where you have implemented them it would be great | 17:24 |
sidsniper | ORL and CBCL are face database right? | 17:26 |
blackburn | sidsniper: implemented what? | 17:26 |
blackburn | sidsniper: yes | 17:26 |
sidsniper | implemented the algorithms | 17:26 |
sidsniper | dim red | 17:27 |
blackburn | hmm I mainly refer to dr toolbox | 17:27 |
blackburn | http://homepage.tudelft.nl/19j49/Matlab_Toolbox_for_Dimensionality_Reduction.html | 17:27 |
sidsniper | nice :) | 17:28 |
sidsniper | i have to implement some of them for my project | 17:28 |
sidsniper | can u suggest how to proceed? | 17:28 |
sidsniper | i mean which are the easier ones | 17:28 |
blackburn | sidsniper: LLE is pretty simple | 17:30 |
blackburn | diffusion maps are as well | 17:30 |
blackburn | they are all actually pretty simple | 17:30 |
blackburn | but pretty slow in case of straightforward implementation | 17:31 |
sidsniper | so implementation part is hard? | 17:31 |
blackburn | depends on speed you need | 17:31 |
blackburn | I had to implement superlu and arpack wrappers | 17:31 |
blackburn | and to parallelize it | 17:31 |
blackburn | but in matlab it could be 20-liner | 17:32 |
sidsniper | okay | 17:32 |
blackburn | as for correctness | 17:33 |
sidsniper | but we have to do it in R language | 17:33 |
blackburn | they are all ill-posed | 17:33 |
sidsniper | will that be tough? | 17:33 |
blackburn | no idea | 17:33 |
sidsniper | i saw that shogub also suppports R | 17:33 |
sidsniper | shogun* | 17:33 |
blackburn | how much data you have? | 17:33 |
-!- harshit_ [~harshit@182.68.249.161] has joined #shogun | 17:33 | |
sidsniper | cureently i dont have | 17:33 |
blackburn | harshit_: congrats your code is merged ;) | 17:34 |
sidsniper | i will first try on swissroll as suggested by u | 17:34 |
blackburn | sidsniper: yeah try ;) | 17:34 |
harshit_ | hey blackburn: is there any way to get precision , recall, accuracy etc on tested dataset in shogun itself ? | 17:34 |
harshit_ | and thanks i didnt check it , | 17:35 |
blackburn | harshit_: yes, ContingencyTableEvaluation | 17:35 |
blackburn | ok heading home finally | 17:35 |
blackburn | see you | 17:36 |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 17:36 | |
harshit_ | github is strange doesn't notify you for getting your code merged ! | 17:36 |
sidsniper | harshit_: which project are u applying for? | 17:37 |
harshit_ | sidsniper :Its a collection of things actually | 17:39 |
gsomix | sonney2k, sonne|work slap. | 17:39 |
harshit_ | 2 projects for which i am applying are: integrate regression in liblinear + implement some cdotfeatures | 17:40 |
harshit_ | and one is trees + liblinear | 17:40 |
harshit_ | sidsniper : what about you ? | 17:40 |
sidsniper | okay | 17:40 |
wiking | dman | 17:47 |
gsomix | sonney2k, sonne|work I want to finish my proposal today. That's all I want. :) | 17:50 |
gsomix | however I can do smth else. | 17:50 |
shogun-buildbot | build #455 of octave_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/455 blamelist: harshit.syal@nsitonline.in | 18:00 |
shogun-buildbot | build #442 of csharp_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/csharp_modular/builds/442 blamelist: harshit.syal@nsitonline.in | 18:04 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 18:08 | |
shogun-buildbot | build #431 of lua_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/lua_modular/builds/431 blamelist: harshit.syal@nsitonline.in | 18:09 |
-!- khayyam [~chatzilla@189.166.217.6] has joined #shogun | 18:11 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 18:13 | |
shogun-buildbot | build #434 of java_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/java_modular/builds/434 blamelist: harshit.syal@nsitonline.in | 18:13 |
harshit_ | n4nd0: hey , do you know why all these build are are failing to compile ? | 18:15 |
harshit_ | As in does shogun-buildbot compiles every commit that i made earlier | 18:15 |
n4nd0 | harshit_: so shogun-buildbot compiles commit by commit | 18:15 |
n4nd0 | harshit_: exactly, can it be that some of those commits where not complete? | 18:16 |
harshit_ | I think there were some commits that I made earlier,in which there were some conflicts | 18:17 |
n4nd0 | I can see this error ^ | 18:17 |
n4nd0 | make[1]: *** No rule to make target `../../shogun/classifier/svm/NewtonSVM.h', needed by `modshogun_wrap.cxx'. Stop. | 18:17 |
harshit_ | So what do I need to do now | 18:17 |
n4nd0 | for me it looks like NewtonSVM.h was missing at that moment from the interfaces | 18:17 |
n4nd0 | but I am not sure | 18:17 |
n4nd0 | harshit_: don't worry about it | 18:17 |
n4nd0 | harshit_: it should compile the following commits soon and recover the state | 18:18 |
harshit_ | ohkay got it | 18:18 |
shogun-buildbot | build #430 of python_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/430 blamelist: harshit.syal@nsitonline.in | 18:18 |
n4nd0 | harshit_: you can monitor it a bit if you are interested | 18:18 |
n4nd0 | http://shogun-toolbox.org/buildbot/waterfall | 18:18 |
n4nd0 | just if you are curious to know how it works | 18:19 |
n4nd0 | I find it quite cool | 18:19 |
n4nd0 | (I think I like everything that is colorful :P) | 18:19 |
harshit_ | ^ dont blame me, :( | 18:20 |
harshit_ | So that was the reason why blackburn said to make big commits rather than small ones | 18:20 |
harshit_ | looks great | 18:21 |
n4nd0 | exactly | 18:21 |
-!- naywhaya1e is now known as naywhayare | 18:21 | |
n4nd0 | there's another way it seems | 18:21 |
n4nd0 | that sonney2k proposes using git commit -amend | 18:22 |
n4nd0 | I don't know how to use it yet though | 18:22 |
harshit_ | what does -amend do ? | 18:22 |
shogun-buildbot | build #430 of r_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_modular/builds/430 blamelist: harshit.syal@nsitonline.in | 18:22 |
n4nd0 | I have not really gone into it yet, but I think it is used to replace already published-commits | 18:24 |
n4nd0 | take a look to man git commit | 18:24 |
harshit_ | oh great, I think we need to use it more often then | 18:24 |
n4nd0 | and search for --amend | 18:24 |
harshit_ | yeah doing the same | 18:25 |
shogun-buildbot | build #426 of ruby_modular is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/426 blamelist: harshit.syal@nsitonline.in | 18:27 |
-!- harshit_ [~harshit@182.68.249.161] has quit [Quit: Leaving] | 18:30 | |
-!- blackburn [~qdrgsm@83.234.54.186] has joined #shogun | 18:34 | |
shogun-buildbot | build #456 of octave_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/octave_modular/builds/456 | 18:35 |
shogun-buildbot | build #443 of csharp_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/csharp_modular/builds/443 | 18:42 |
n4nd0 | great, our friend shogun-buildbot is healthy again | 18:42 |
shogun-buildbot | build #432 of lua_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/lua_modular/builds/432 | 18:48 |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 18:48 | |
shogun-buildbot | build #435 of java_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/java_modular/builds/435 | 18:58 |
-!- flxb [~cronor@fb.ml.tu-berlin.de] has quit [Quit: flxb] | 18:59 | |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Ping timeout: 276 seconds] | 19:02 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 19:03 | |
PhilTillet | Hey!! | 19:05 |
PhilTillet | :) | 19:05 |
n4nd0 | hi! | 19:05 |
shogun-buildbot | build #431 of python_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/python_modular/builds/431 | 19:06 |
PhilTillet | sup? | 19:06 |
n4nd0 | bah not much | 19:06 |
n4nd0 | what about you? | 19:06 |
n4nd0 | our friendly shogun-buildbot is almost recovered again :) | 19:07 |
PhilTillet | lol :D | 19:07 |
PhilTillet | well not much | 19:07 |
PhilTillet | i'm tired | 19:08 |
PhilTillet | and hungry | 19:08 |
n4nd0 | dinner time? | 19:09 |
PhilTillet | well | 19:09 |
PhilTillet | sort of | 19:09 |
PhilTillet | but can't have dinner | 19:09 |
PhilTillet | the canteen is not opened yet | 19:10 |
n4nd0 | aham | 19:11 |
shogun-buildbot | build #431 of r_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/r_modular/builds/431 | 19:12 |
-!- harshit_ [~harshit@182.68.249.161] has joined #shogun | 19:15 | |
shogun-buildbot | build #427 of ruby_modular is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/ruby_modular/builds/427 | 19:19 |
-!- Juggy [~Razor@14.139.82.6] has joined #shogun | 19:20 | |
Juggy | Hi blackburn | 19:21 |
Juggy | Can you give me the link for tutorial on Dimensionality Reduction in shogun? | 19:21 |
Juggy | @blackburn: are you there ? | 19:33 |
-!- harshit_ [~harshit@182.68.249.161] has quit [Ping timeout: 252 seconds] | 19:40 | |
-!- harshit_ [~harshit@182.68.249.161] has joined #shogun | 19:44 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Ping timeout: 264 seconds] | 19:46 | |
@sonney2k | harshit_, I am doing some whitespace cleanups and will commit soon - please have a brief look then. I hope this helps to make clear what I had in mind coding style wise | 19:46 |
n4nd0 | sonney2k: I am sorry to insist but, did you have time to take a look at multiclass proposal? | 19:47 |
gsomix | sonney2k, moin | 19:50 |
@sonney2k | n4nd0, sorry not yet | 19:50 |
@sonney2k | gsomix, so how is it going? | 19:50 |
n4nd0 | sonney2k: all right, I am sorry that I rushed | 19:51 |
CIA-64 | shogun: Soeren Sonnenburg master * rcfab94a / (2 files): whitespace and coding style fixed in newtonsvm - http://git.io/wnQ-rw | 19:52 |
gsomix | sonney2k, i'm working on my proposal and python3 interface now. | 19:54 |
gsomix | sonney2k, are there other tasks for me? | 19:55 |
@sonney2k | harshit_, please have a look at http://git.io/wnQ-rw | 19:56 |
-!- puneetgoyal [~puneet@115.242.16.126] has joined #shogun | 19:56 | |
@sonney2k | gsomix, most important now is to write your proposal | 19:57 |
@sonney2k | gsomix, apart from that it depends a bit what you are most interested in doing | 20:00 |
@sonney2k | I guess python3 stuff is close to final, so other simple task is to write swig typemap for octave modular | 20:00 |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has joined #shogun | 20:00 | |
harshit_ | sonney2k : sorry having issues with Internet | 20:01 |
@sonney2k | yeah I've seen... | 20:01 |
@sonney2k | harshit_, btw can you write some more description about newton svm in the class description and mention the paper there? | 20:02 |
@sonney2k | harshit_, ahh and before I forget - you didn't mention how shogun's newtonsvm performs on covertype data set vs. matlab version | 20:03 |
@sonney2k | would be interesting to know | 20:03 |
harshit_ | sure, i'll do it today | 20:03 |
gsomix | sonney2k, ok. i will consider this issue later. | 20:04 |
PhilTillet | sonney2k, what are the "most used" CKernel children? DotKernel I suppose, but any other? :p | 20:04 |
harshit_ | and for comparison there was a problem running the dataset on matlab version, so i compared it with liblinear results on that dataset | 20:04 |
harshit_ | liblinear took about 3-4 sec to train vs my algo which took about 6-8 sec | 20:05 |
harshit_ | and the final weight matrix was almost similar | 20:05 |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 20:05 | |
@sonney2k | gsomix, please ask questions / update us - I am not really sure if the tasks you have are too boring for you. | 20:06 |
@sonney2k | PhilTillet, the most used kernel is the gaussian kernel | 20:06 |
harshit_ | sonney2k : so after inclusion of documentation in NewtonSVM.cpp, do i need to issue a new pull request ? | 20:06 |
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 20:06 | |
@sonney2k | harshit_, sure | 20:06 |
@sonney2k | PhilTillet_, the most used kernel is the gaussian kernel | 20:07 |
@sonney2k | harshit_, what was the problem on matlab? | 20:07 |
PhilTillet_ | yes, Actually every DotKernel could be more or less implemented the same way on OpenCL | 20:07 |
PhilTillet_ | but might be different for some other kernel classes | 20:07 |
PhilTillet_ | :p | 20:07 |
PhilTillet_ | brb eat :D | 20:07 |
harshit_ | sonney2k: matlab said there is some problem with minres() function when i used that dataset | 20:08 |
harshit_ | but otherwise with other dataset it ran perfectly | 20:08 |
-!- PhilTillet [~Philippe@npasserelle10.minet.net] has quit [Ping timeout: 250 seconds] | 20:09 | |
@sonney2k | harshit_, and if you use a subset of this data set (say first 10000 examples only?) | 20:09 |
harshit_ | I think that dataset is pretty big | 20:09 |
harshit_ | to handle | 20:09 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Remote host closed the connection] | 20:09 | |
harshit_ | oops forgot to do that ! | 20:09 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 20:09 | |
@sonney2k | PhilTillet_, dotkernel is most important... / maybe a function to compute whole kernel/distance matrix based on opencl would be nice too | 20:10 |
harshit_ | I'll do it today and report you, for now preparing my gsoc proposal ( i think i am running late on that) | 20:10 |
@sonney2k | harshit_, yes that is most important true | 20:10 |
harshit_ | sonney2k: after submitting my proposal i want to start working on lbp features implementation . | 20:11 |
harshit_ | do you think opencv's implementation is portable ? | 20:12 |
@sonney2k | harshit_, we have one in shogun that needs fixing - matlab / c code is avilable | 20:13 |
harshit_ | sonney2k: LBP features is already in shogun ! I thought that needs to be implemented from starting | 20:15 |
@sonney2k | harshit_, it is in unfinished state | 20:16 |
harshit_ | anyways that is great I can start working on it ~ | 20:16 |
harshit_ | please just give me any link to the matlab/c code from which it was ported | 20:17 |
@sonney2k | harshit_, when it is in you can do some nice demo using opencv and these LBP features - to detect gender of a person based on an image of the face | 20:17 |
harshit_ | sonney2k : actually I have done something like that using deep learning "sparse autoencoders" | 20:18 |
harshit_ | I think it would be easy for me | 20:18 |
harshit_ | great thought | 20:18 |
@sonney2k | karlnapf, around? | 20:20 |
gsomix | sonney2k, boring problem - this is also the problem. :) | 20:20 |
karlnapf | sonney2k, yes hi | 20:20 |
karlnapf | whats up? | 20:20 |
@sonney2k | karlnapf, would you mind if I give you github commit rights - I mean you are a core developer without this right for some time now... | 20:21 |
@sonney2k | you can still do the same workflow | 20:21 |
karlnapf | why should I mind this? :) | 20:21 |
karlnapf | Should I have a look at people's requests then? | 20:21 |
@sonney2k | karlnapf, I knew that you like to break things | 20:22 |
@sonney2k | karlnapf, if you have time - that's welcome | 20:22 |
karlnapf | sure, of course then | 20:22 |
karlnapf | I will still ask you guys to take a look on stuff which involves larger changes | 20:22 |
karlnapf | sonney2k, I am just wondering what is our favorite way of doing intending in python, spaces or tabs? | 20:24 |
blackburn | Juggy: still there? | 20:24 |
blackburn | sonney2k: finally! | 20:25 |
blackburn | Juggy: http://shogun-toolbox.org/edrt/tutorial.pdf probably that one, it is unaccurate though | 20:25 |
blackburn | in* | 20:26 |
blackburn | PhilTillet_: btw I have something really big to propose | 20:28 |
blackburn | I have some collection of dimensionality reduction algorithms there.. ;) | 20:29 |
@sonney2k | karlnapf, https://github.com/shogun-toolbox - welcome to the team but now officially :D | 20:29 |
karlnapf | nice one :D | 20:29 |
karlnapf | thanks | 20:29 |
blackburn | sonney2k: what a long way home | 20:29 |
-!- gsomix [~gsomix@188.168.5.227] has quit [Read error: Operation timed out] | 20:29 | |
@sonney2k | karlnapf, well *I* do tabs as usual | 20:29 |
blackburn | I prefer tabs as well btw | 20:30 |
@sonney2k | blackburn, sing some russian love song for heiko please - to celebrate this event! | 20:30 |
karlnapf | yeaaah :) | 20:30 |
blackburn | sonney2k: love song? like 'it is ok to be gay'? | 20:30 |
* sonney2k turns the volume to the max! | 20:30 | |
karlnapf | blackburn, gogo | 20:30 |
@sonney2k | blackburn, no sth. like ~~~oh~~~heiko~~~in~~~your~~~code~~~ohh~~~we~~~trust~~~ | 20:31 |
blackburn | ??????????? ???? ??????? ???? ?????? | 20:31 |
@sonney2k | karlnapf, regarding spaces - it might be that python3 wants spaces ... not sure | 20:31 |
blackburn | ? ?????????? ? ????????? ???? ????? | 20:31 |
karlnapf | sing louder blackburn, cant hear you | 20:31 |
blackburn | ???? ??????? ?? ???? ???????? | 20:31 |
blackburn | ?? ??????????? ????????????????????????! | 20:31 |
karlnapf | sonney2k, ok, but we can convert, just asking | 20:32 |
blackburn | kind of old love song hahaha | 20:32 |
blackburn | just like sonney2k asked | 20:32 |
karlnapf | I love these letters | 20:32 |
karlnapf | want to learn Russian at some point | 20:32 |
blackburn | funny it is a war song probably hahah | 20:32 |
blackburn | WW2 yes | 20:32 |
karlnapf | Ill double check with google translate | 20:32 |
blackburn | karlnapf: would be pretty senseless | 20:32 |
karlnapf | Gr?ne Ahorn Riegelahorn, aber ;) | 20:33 |
@sonney2k | *rotflbtc* | 20:33 |
karlnapf | but I can heart it, the friendly google woman sings it | 20:33 |
blackburn | http://www.youtube.com/watch?feature=player_detailpage&v=CXMEto5NXb8 | 20:34 |
blackburn | karlnapf: this one is better hah | 20:34 |
-!- sidsniper [~piggy@14.139.82.6] has quit [Ping timeout: 250 seconds] | 20:34 | |
karlnapf | nice | 20:34 |
blackburn | karlnapf: why do you want to learn russian? | 20:35 |
karlnapf | Want to learn another language which is not romanian | 20:35 |
karlnapf | and I want to travel to the place | 20:35 |
blackburn | karlnapf: you are very welcome btw ;) | 20:36 |
-!- Juggy [~Razor@14.139.82.6] has quit [Ping timeout: 260 seconds] | 20:36 | |
karlnapf | thanks blackburn | 20:36 |
karlnapf | its very nice to be in your team guys | 20:36 |
blackburn | karlnapf: seriously I would like someone of you to visit me here | 20:36 |
karlnapf | I was also serious, really wanna go there, so perhaps next year or so, well meet | 20:37 |
blackburn | sonney2k: let you organize shogun code sprint at berlin? ;) | 20:37 |
karlnapf | also a nice idea | 20:38 |
-!- gsomix [~gsomix@188.168.2.47] has joined #shogun | 20:38 | |
@sonney2k | at my flat with my kids jumping around - why not | 20:40 |
blackburn | how big is your flat? ;) | 20:40 |
@sonney2k | no but seriously we could do that from gsoc money | 20:40 |
karlnapf | hehe, well I know many people to stay at | 20:40 |
karlnapf | yes would be kind of cool to meet | 20:41 |
karlnapf | blackburn, how is your german? :) | 20:41 |
blackburn | karlnapf: ~ to your russian ;) | 20:41 |
blackburn | ich bin mude | 20:42 |
@sonney2k | ich bin immer muede | 20:42 |
blackburn | karlnapf: wo bist du? | 20:42 |
karlnapf | ich bin in england | 20:42 |
blackburn | sonney2k: ??? ????? | 20:42 |
karlnapf | ich bin sehr m?de | 20:42 |
karlnapf | ???? | 20:43 |
blackburn | yes I know umlaut is there | 20:43 |
blackburn | too lazy to type it | 20:43 |
blackburn | ? | 20:43 |
karlnapf | how to type it on russian keyboard? | 20:44 |
blackburn | karlnapf: do I remember that correctly - u mit dem umlaut = ? | 20:44 |
blackburn | ? | 20:44 |
blackburn | using compose key | 20:44 |
blackburn | u+right alt+" | 20:44 |
karlnapf | yes | 20:44 |
blackburn | karlnapf: I had been studying german up to 7th grade at school | 20:45 |
@sonney2k | another very strong proposal got submitted... many are from phd students this time | 20:45 |
blackburn | for two years probably | 20:45 |
blackburn | sonney2k: w/o contributions? | 20:45 |
karlnapf | ah really, thats more than my russian | 20:45 |
karlnapf | which project? | 20:45 |
@sonney2k | libqp | 20:45 |
gsomix | >> phd students *__* | 20:45 |
karlnapf | yes, phd all over the place, gsoc is very academic | 20:46 |
blackburn | my belief is proposal w/o contributions would not work | 20:46 |
@sonney2k | karlnapf, last year that was not the case | 20:47 |
karlnapf | sonney2k, probably means that stuff is more advanced, or they checked last years projects or so | 20:47 |
karlnapf | I mean if you dont do machine learning the stuff is pretty tough | 20:48 |
karlnapf | even if you do its not trivial | 20:48 |
@sonney2k | many things this year are really tough | 20:48 |
@sonney2k | for example for SO we didn't have any proposal last time | 20:48 |
@sonney2k | this year quite a bit | 20:48 |
blackburn | this year we have mighty n4nd0 | 20:49 |
karlnapf | SO? | 20:49 |
blackburn | structured ouput | 20:49 |
blackburn | tput | 20:49 |
karlnapf | ah ok | 20:49 |
karlnapf | yes, cool thing is that | 20:49 |
blackburn | n4nd0: btw checking your spe right now | 20:50 |
@sonney2k | anyway we have 9 mentors so max slots is 9 | 20:52 |
karlnapf | lets hope to get all that ! | 20:52 |
blackburn | uh | 20:52 |
gsomix | oh | 20:52 |
blackburn | pretty impossible probably | 20:52 |
@sonney2k | karlnapf, actually in retrospect I am happy that we got just 5 slots last year | 20:53 |
@sonney2k | it really was a lot of work | 20:53 |
karlnapf | sonney2k, ok I believe that | 20:53 |
@sonney2k | this year though we have n4nd0, blackburn, karlnapf and others helping others | 20:53 |
@sonney2k | it is not soo much all on my shoulders any more | 20:53 |
karlnapf | yes, you did almost everythign last year | 20:53 |
@sonney2k | so maybe this process can scale | 20:53 |
@sonney2k | and actually students this year all submitted github pull requests before | 20:54 |
@sonney2k | with pretty nice work that goes deep into shogun | 20:54 |
@sonney2k | last year we were switching from svn -> git in the process | 20:55 |
karlnapf | this year we got the buildbots | 20:55 |
@sonney2k | we were doing conversions like double* -> sgvector | 20:55 |
@sonney2k | we created buildbots | 20:55 |
@sonney2k | etc | 20:55 |
-!- flxb [~cronor@g231218093.adsl.alicedsl.de] has joined #shogun | 20:55 | |
@sonney2k | lots of infrastructure that is now there and helps to keep the project in a sane state | 20:55 |
@sonney2k | to conclude - we should be realistically able to handle 9 students | 20:56 |
blackburn | sonney2k: what was there before gsoc 2011? | 20:56 |
karlnapf | sonney2k, wow that would be amazing. | 20:56 |
karlnapf | do you know how google allocates slots? | 20:56 |
karlnapf | only in terms of application # | 20:56 |
karlnapf | or also other stuff? | 20:56 |
blackburn | karlnapf: no, there are some criteria | 20:58 |
blackburn | I'll refer in a min | 20:58 |
blackburn | karlnapf: http://code.google.com/p/google-summer-of-code/wiki/NotesonStudentAllocations | 20:59 |
-!- Juggy [~Razor@14.139.82.6] has joined #shogun | 20:59 | |
-!- sidsniper [~piggy@14.139.82.6] has joined #shogun | 20:59 | |
n4nd0 | blackburn: cool man, I was having dinner, tell me if I should do / say sth | 21:01 |
blackburn | karlnapf: to be fair we fit 1), 2), 4) | 21:01 |
karlnapf | sonney2k, just trying out a request | 21:01 |
karlnapf | does the workflow change? | 21:01 |
karlnapf | do I still make pull requests and then merge them myself? | 21:01 |
blackburn | n4nd0: I like it pretty much | 21:01 |
blackburn | karlnapf: I think this type of commits would be better to be direct | 21:02 |
karlnapf | ok so only pull request for larger stuff | 21:02 |
karlnapf | and direct means I clone the actual shogun repo? | 21:03 |
blackburn | karlnapf: yes | 21:03 |
blackburn | karlnapf: ah yes | 21:03 |
blackburn | you need to push to | 21:03 |
karlnapf | ok then | 21:03 |
blackburn | google, shogun and github | 21:04 |
@sonney2k | karlnapf, btw you can from now on just do (C) Heiko S... no need to mention TU or so any longer | 21:04 |
blackburn | sonney2k: btw | 21:04 |
blackburn | I had idea before | 21:04 |
blackburn | Written and Copyright now are pretty same | 21:04 |
@sonney2k | blackburn, karlnapf I disagree | 21:04 |
blackburn | sonney2k: about? | 21:04 |
@sonney2k | I am very much favoring pull requests | 21:05 |
@sonney2k | and then have someone else looking at it | 21:05 |
blackburn | sonney2k: for small fixes? | 21:05 |
@sonney2k | but yeah I know blackburn you disagree | 21:05 |
@sonney2k | blackburn, for small fixes pull request + self merge | 21:05 |
blackburn | this takes more time.. | 21:05 |
karlnapf | ok like examples and stuff, but its always good to get stuff reviewed when its more complicated | 21:06 |
@sonney2k | and avoids quite a bit of stupid errors | 21:06 |
blackburn | sonney2k: these stupid errors emerge pretty rarely.. | 21:06 |
karlnapf | blackburn, I have a talent of finding stuff like this ;) | 21:07 |
blackburn | I still don't like this kind of self-merge | 21:07 |
CIA-64 | shogun: Heiko Strathmann master * rf5c30fd / examples/undocumented/python_modular/modelselection_grid_search_kernel.py : added example for modelselection of different kernels - http://git.io/y1hxQA | 21:07 |
CIA-64 | shogun: Heiko Strathmann master * r67c3782 / examples/undocumented/python_modular/modelselection_grid_search_kernel.py : fixed copyright and excluded TUB - http://git.io/2pNQHw | 21:07 |
CIA-64 | shogun: Heiko Strathmann master * r2801f8d / examples/undocumented/python_modular/modelselection_grid_search_kernel.py : | 21:07 |
CIA-64 | shogun: Merge pull request #417 from karlnapf/master | 21:07 |
CIA-64 | shogun: new example - http://git.io/ehsJzw | 21:07 |
blackburn | karlnapf: sonney2k: one more thing I would like to suggest | 21:08 |
blackburn | is to use #XXX stuff | 21:08 |
blackburn | for example if you are pretty ready to merge PR no need to wait | 21:08 |
blackburn | you may commit with 'blabla Closes #XXX' | 21:09 |
blackburn | and actually we can use not only PRs but issues as well | 21:09 |
blackburn | we can commit with #XXX to incorporate commit to this issue | 21:10 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has quit [Ping timeout: 276 seconds] | 21:10 | |
blackburn | I believe it makes more sense to group commits with issues | 21:10 |
@sonney2k | ...if we used issues | 21:10 |
blackburn | let me show how it works | 21:10 |
-!- muddo [~muddo@gateway/tor-sasl/muddo] has joined #shogun | 21:11 | |
blackburn | https://github.com/shogun-toolbox/shogun/issues/418 | 21:11 |
shogun-buildbot | build #198 of nightly_none is complete: Failure [failed compile] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_none/builds/198 | 21:11 |
blackburn | here we would have docu improvements | 21:11 |
n4nd0 | karlnapf: about the answer to the issue, can you elaborate a bit more on that? what is the idea then? not to serialize SGNArray stuff? | 21:13 |
n4nd0 | karlnapf: I think that might be a problem because for example QDA has some members of that type | 21:13 |
karlnapf | n4nd0, no that was just a general thought | 21:13 |
karlnapf | this is possible (and very nice I think) | 21:13 |
karlnapf | but its more complicated than on the first impression | 21:13 |
blackburn | and once I add new commit | 21:14 |
n4nd0 | karlnapf: I see, I have no idea how serialization stuff works in shogun | 21:14 |
karlnapf | if you dive into the load_serializable method of CSGObject you will that there are many places where these types have to be handled | 21:14 |
karlnapf | n4nd0, is it important that you need that? | 21:15 |
CIA-64 | shogun: Sergey Lisitsyn master * rb1b17cb / src/shogun/converter/DiffusionMaps.h : Added note about lapack performance with diffusion maps #418 - http://git.io/4zEjlQ | 21:15 |
blackburn | it appears in #418 | 21:15 |
blackburn | https://github.com/shogun-toolbox/shogun/issues/418 | 21:15 |
n4nd0 | karlnapf: you mean if I need to use that type in QDA? | 21:15 |
blackburn | sonney2k: karlnapf: what do you think? | 21:15 |
karlnapf | n4nd0, yes | 21:16 |
n4nd0 | karlnapf: at the beginning I was simply using SGMatrix * | 21:16 |
n4nd0 | karlnapf: the idea is that I need a list of matrices | 21:16 |
karlnapf | blackburn, I think thats nice, but is it easy to handle? | 21:16 |
n4nd0 | karlnapf: but that SGMatrix* turned out to be a bit problematic also with this issue, the serialization | 21:17 |
blackburn | karlnapf: yes all you need is to recall # of issue | 21:17 |
karlnapf | n4nd0, i see | 21:17 |
n4nd0 | karlnapf: so sonney2k proposed to use SGNDArray instead | 21:17 |
@sonney2k | blackburn, yeah I like that - we had this same feature with trac before but also there rarely used it - just because we never really managed to use issues all the time | 21:17 |
blackburn | sonney2k: that would be faster for doc fixes and other small fixes | 21:17 |
blackburn | too much pull requests would slow down things I think | 21:18 |
karlnapf | n4nd0, ok, probably good idea then | 21:18 |
@sonney2k | blackburn, I am saying lets use it but I don't know how who will file issues all the time | 21:18 |
@sonney2k | not sure if we are enough developers yet for that stuff | 21:18 |
karlnapf | n4nd0, I just hate to touch the migration stuff ;), but should be doable, if you want, take a look | 21:18 |
n4nd0 | karlnapf: so the thing is that right now serialization in QDA wouldn't work right? | 21:18 |
karlnapf | n4nd0, yes | 21:19 |
blackburn | sonney2k: hmm just write somewhere (doc fixes goes to #418) | 21:19 |
karlnapf | not too bad now, but should be done on the long run | 21:19 |
n4nd0 | karlnapf: I will write that in the doc. at least | 21:19 |
blackburn | would be nice to make some other keyword referencing | 21:19 |
n4nd0 | karlnapf: yeah, I agree, sooner or later it must be done | 21:19 |
@sonney2k | blackburn, can I close an issue from via a commit message? | 21:19 |
karlnapf | n4nd0, yes, perhaps during the summer :) | 21:20 |
blackburn | sonney2k: closes #418 | 21:20 |
@sonney2k | k | 21:20 |
blackburn | sonney2k: just like pull request | 21:20 |
blackburn | but no need to switch between fork and origin | 21:20 |
blackburn | sonney2k: fixes, fixed, fix, closes, close, closed | 21:21 |
blackburn | and of these words | 21:21 |
blackburn | any* | 21:21 |
blackburn | ok have to clean up my damned traffic sign recognition paper be back in a hour or so | 21:24 |
gsomix | sonney2k, http://pastebin.com/DuSe6Qs1 at now. my last fixes and 2to3. | 21:24 |
n4nd0 | blackburn: see you later then, maybe we can finish with SPE then? | 21:24 |
blackburn | n4nd0: hmm one more thing | 21:24 |
blackburn | please add reference | 21:25 |
blackburn | and I'll check it once again this night and merge | 21:25 |
n4nd0 | blackburn: for the doc? | 21:25 |
blackburn | yes | 21:25 |
n4nd0 | blackburn: ok | 21:25 |
-!- khayyam [~chatzilla@189.166.217.6] has quit [Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120312181643]] | 21:25 | |
n4nd0 | blackburn: I am a bad guy ... I am always like saying TODO doc :P | 21:25 |
blackburn | heh better have todo than forgot | 21:25 |
@sonney2k | gsomix, you need to install liblzma-dev | 21:26 |
-!- blackburn [~qdrgsm@83.234.54.186] has quit [Quit: Leaving.] | 21:26 | |
@sonney2k | gsomix, and libjson0-dev | 21:26 |
gsomix | sonney2k, done | 21:31 |
sidsniper | blackburn can you give me the link to the tutorial on dim red? | 21:36 |
n4nd0 | sidsniper: I think blackburn is out now working on his stuff | 21:37 |
n4nd0 | sidsniper: what tutorial do you want? | 21:37 |
sidsniper | the one that is in shogun | 21:38 |
sidsniper | tutorial showing how to use those algos | 21:38 |
n4nd0 | ok, have you taken a look to the examples? | 21:38 |
n4nd0 | maybe that is enough for your needs? | 21:38 |
gsomix | sonney2k, but doesn't work. hmm | 21:40 |
gsomix | on python3. | 21:40 |
gsomix | and on python2. | 21:41 |
gsomix | reboot | 21:41 |
-!- gsomix [~gsomix@188.168.2.47] has quit [Quit: ????? ? ?? ??? (xchat 2.4.5 ??? ??????)] | 21:41 | |
-!- Afan [~Afan@c-76-29-50-209.hsd1.il.comcast.net] has joined #shogun | 21:48 | |
PhilTillet_ | Hello again | 21:48 |
PhilTillet_ | :) | 21:48 |
-!- PhilTillet_ is now known as PhilTillet | 21:51 | |
-!- gsomix [~gsomix@83.234.169.122] has joined #shogun | 21:52 | |
gsomix | sonney2k, works. | 21:53 |
-!- uricamic [5546c513@gateway/web/freenode/ip.85.70.197.19] has joined #shogun | 21:53 | |
Juggy | is there who knows about the DR tutorial in Shogun? | 22:01 |
Juggy | I am not able to fine :( | 22:01 |
Juggy | *find | 22:01 |
Juggy | is there a list of implemented algorithms ????? | 22:02 |
n4nd0 | you can check them under the directory converter | 22:03 |
n4nd0 | http://www.shogun-toolbox.org/doc/en/current/classshogun_1_1CConverter.html | 22:04 |
n4nd0 | there you can a find a list in doxygen doc | 22:04 |
n4nd0 | but doc might not be up to date | 22:05 |
@sonney2k | n4nd0, strong MC proposal... | 22:07 |
@sonney2k | yours I mean | 22:08 |
n4nd0 | sonney2k: oh thank you :) | 22:08 |
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has joined #shogun | 22:08 | |
-!- PhilTillet [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Remote host closed the connection] | 22:08 | |
-!- gsomix [~gsomix@83.234.169.122] has quit [Quit: ????? ? ?? ??? (xchat 2.4.5 ??? ??????)] | 22:21 | |
PhilTillet_ | sonney2k, do you have any idea of what "big idea" blackburn was talking about? | 22:38 |
@sonney2k | big idea? | 22:47 |
@sonney2k | I guess I missed sth | 22:47 |
@sonney2k | except he is dreaming of world domination again ;-) | 22:47 |
PhilTillet_ | <blackburn> PhilTillet_: btw I have something really big to propose | 22:49 |
PhilTillet_ | <blackburn> I have some collection of dimensionality reduction algorithms there.. ;) | 22:49 |
PhilTillet_ | XD | 22:50 |
@sonney2k | many dim reduction methods are expensive so it sounds like he wants some opencl magic :) | 22:51 |
PhilTillet_ | haha | 22:51 |
@sonney2k | PhilTillet_, btw how are you progressing with your patch? | 22:51 |
PhilTillet_ | They are all about eigenvalues right? | 22:51 |
PhilTillet_ | Well, pretty hard | 22:51 |
PhilTillet_ | I cleaned my source | 22:52 |
PhilTillet_ | so that there is a clear OpenCL structure | 22:52 |
PhilTillet_ | to compile programs etc | 22:52 |
PhilTillet_ | and now I copy with direct memcopy, without using ublas | 22:52 |
PhilTillet_ | I mean, for now my patch is mostly the opencl code | 22:54 |
PhilTillet_ | and basically a shogun::ocl::svm::init() | 22:54 |
PhilTillet_ | which compiles all the opencl code | 22:54 |
-!- wiking [~wiking@huwico/staff/wiking] has quit [Quit: wiking] | 23:03 | |
PhilTillet_ | memory management is one of the first thing in my GSoC Timeline, but won't be able to do it for the patch | 23:04 |
PhilTillet_ | :D | 23:04 |
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has joined #shogun | 23:18 | |
-!- wiking [~wiking@78-23-191-201.access.telenet.be] has quit [Changing host] | 23:18 | |
-!- wiking [~wiking@huwico/staff/wiking] has joined #shogun | 23:18 | |
-!- Marty28 [~Marty@158.181.76.57] has quit [Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120310010446]] | 23:18 | |
-!- karlnapf [~heiko@host86-181-9-49.range86-181.btcentralplus.com] has quit [Quit: Leaving.] | 23:22 | |
PhilTillet_ | sonney2k, hope I'll be able to submit an opencl patch (not perfect from a design point of view XD) soon | 23:28 |
CIA-64 | shogun: Soeren Sonnenburg master * r21ae0ec / src/shogun/classifier/svm/NewtonSVM.h : add ifdef HAVE_LAPACK for NewtonSVM - http://git.io/iX7SlQ | 23:31 |
-!- puneetgoyal [~puneet@115.242.16.126] has quit [Ping timeout: 246 seconds] | 23:32 | |
PhilTillet_ | sonney2k, I'll have some design question though :p | 23:36 |
-!- uricamic [5546c513@gateway/web/freenode/ip.85.70.197.19] has quit [Quit: Page closed] | 23:51 | |
-!- PhilTillet_ [~Philippe@tillet-p42154.maisel.int-evry.fr] has quit [Quit: Leaving] | 23:57 | |
-!- harshit_ [~harshit@182.68.249.161] has quit [Ping timeout: 244 seconds] | 23:57 | |
--- Log closed Wed Apr 04 00:00:19 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!