--- Log opened Mon Jan 28 00:00:38 2013 | ||
-!- FSCV [~FSCV@189.139.229.7] has joined #shogun | 01:04 | |
shogun-buildbot | build #264 of nightly_default is complete: Failure [failed test] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/264 | 03:50 |
---|---|---|
-!- FSCV [~FSCV@189.139.229.7] has quit [Quit: Leaving] | 04:00 | |
-!- cavalier- [cavalier-@70.118.110.73] has quit [Ping timeout: 256 seconds] | 05:53 | |
-!- cavalier- [~cavalier-@70.118.110.73] has joined #shogun | 06:12 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 09:05 | |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun | 09:56 | |
sonne|work | blackburn: how do I add a field to the form? I did add an email addr column | 10:54 |
sonne|work | but that is not automagically in the form | 10:54 |
blackburn | sonne|work: noo, use the form in menu | 10:54 |
blackburn | form (0) | 10:55 |
sonne|work | and then? | 10:55 |
blackburn | edit form | 10:55 |
sonne|work | yes but how do I add it there? | 10:55 |
blackburn | add item in the top left | 10:55 |
sonne|work | I want name and then email address | 10:55 |
sonne|work | and then what? | 10:55 |
blackburn | some field will appear | 10:55 |
blackburn | drag it to the place you want | 10:55 |
blackburn | and edit | 10:55 |
sonne|work | ok maybe I did it correctly | 10:59 |
sonne|work | now | 10:59 |
sonne|work | n4nd0: anything missing? | 10:59 |
n4nd0 | sonne|work: name and email is enough I think | 11:01 |
sonne|work | n4nd0: well look at what blackburn sent around | 11:01 |
n4nd0 | when? I didn't get anything during the past minutes | 11:02 |
sonne|work | blackburn: should we ask about vegeterian food? | 11:02 |
sonne|work | n4nd0: yesterday | 11:02 |
n4nd0 | the google docs, checking atm | 11:03 |
blackburn | I don't know - is it highly relevant? I get to know only one vegetarian so far :D | 11:03 |
sonne|work | blackburn: you barbarians! | 11:04 |
blackburn | sonne|work: how can one do not eat meat :D | 11:04 |
n4nd0 | nothing else I can come up with mmm | 11:05 |
sonne|work | blackburn: do we need badges? | 11:08 |
blackburn | sonne|work: would be cool I think | 11:08 |
blackburn | it could be kind of souvenir :D | 11:09 |
n4nd0 | +1 to the badges idea :) | 11:18 |
sonne|work | t-shirts too right ;^) | 11:20 |
n4nd0 | sonne|work: yeah! :D | 11:27 |
-!- heiko [~heiko@b0f96dbb.bb.sky.com] has joined #shogun | 11:28 | |
sonne|work | heiko: moin! | 11:31 |
sonne|work | heiko: please also give feedback about the form! | 11:31 |
heiko | sonne|work, hi! | 11:31 |
heiko | checking it out ... | 11:31 |
heiko | sonne|work this form is for before the workshop right? | 11:51 |
n4nd0 | heiko: I think is to keep track of the pople who wants to attent to the workshop | 12:02 |
sonne|work | heiko: to register for the workshop | 12:02 |
n4nd0 | so yeah before the workshop | 12:02 |
heiko | ok | 12:02 |
heiko | then I did not understand the question | 12:02 |
heiko | If you weren't satisfied with something, could you explain here what it was? | 12:02 |
heiko | and the rest is in the comments | 12:02 |
sonne|work | heiko: about shogun? | 12:02 |
heiko | but this looks good | 12:02 |
heiko | I see | 12:02 |
heiko | maybe we should formulate this differently | 12:03 |
n4nd0 | what is this lookalike contest that was mentioned the other day about? | 12:03 |
blackburn | heiko: you can edit too ;) | 12:04 |
heiko | blackburn, I see, but I rather comment and let you edit to avoid confusion | 12:04 |
sonne|work | blackburn: once this is done - will this form send an email to the list or do you have to do that? | 12:10 |
blackburn | sonne|work: we can set up notifications | 12:11 |
blackburn | argh | 12:11 |
blackburn | well one can set up notifications for himself | 12:12 |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed] | 13:33 | |
-!- heiko [~heiko@b0f96dbb.bb.sky.com] has quit [Quit: Leaving.] | 13:37 | |
-!- FSCV [~FSCV@187.210.54.165] has joined #shogun | 14:07 | |
-!- FSCV_ [~FSCV@204.45.132.11] has joined #shogun | 14:36 | |
sonne|work | blackburn: well that would work | 14:38 |
sonne|work | I could add some rule matching the pattern to forward this to the mailinglist | 14:38 |
blackburn | sonne|work: I can receive mails and just answer with CC | 14:39 |
sonne|work | yeah but that is :/ | 14:39 |
sonne|work | better directly post it to the list | 14:39 |
-!- FSCV [~FSCV@187.210.54.165] has quit [Ping timeout: 276 seconds] | 14:39 | |
blackburn | I have no idea what is the best way | 14:41 |
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has joined #shogun | 14:48 | |
heiko | blackburn, I dont get you with check possible answers in form. | 14:54 |
blackburn | heiko: ehm have you seen the form? | 14:55 |
heiko | arhg | 14:55 |
heiko | I thought this was the template for it | 14:55 |
blackburn | I suspect you saw only table :D | 14:55 |
heiko | no | 14:55 |
heiko | is it on the website? | 14:55 |
blackburn | heiko: https://docs.google.com/spreadsheet/viewform?formkey=dEFoS0Q2U19qYkdjT2VReFNvNDBoclE6MQ | 14:55 |
blackburn | no, use Form (0) - Edit form in menu | 14:55 |
heiko | I see | 14:56 |
heiko | can I edit? | 14:56 |
blackburn | yes | 14:56 |
heiko | ah cool | 14:57 |
heiko | I did not even know that google docs could do this :) | 14:57 |
blackburn | hah | 14:57 |
blackburn | that would not be cool to ask users to fill in a table | 14:57 |
blackburn | :D | 14:57 |
heiko | no | 15:00 |
heiko | I thought this was just brainstorming, whatever | 15:00 |
heiko | I think we should have this "how to set-up shogun" tutorial, but it finds in no category | 15:00 |
blackburn | heiko: may be in discuss part? | 15:01 |
heiko | its not really a discussion, is it? | 15:01 |
blackburn | I don't know | 15:02 |
heiko | I think this list of ML stuff might be renamed: On which of these topics would you like a tutorial | 15:02 |
heiko | then we could put that in | 15:02 |
blackburn | heiko: feel free to edit I have poor grammar :D | 15:03 |
heiko | done :) | 15:04 |
heiko | ok gotta work now, see you later | 15:05 |
blackburn | see you | 15:05 |
-!- heiko [~heiko@pat-191-250.internal.eduroam.ucl.ac.uk] has quit [Read error: Operation timed out] | 15:44 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has joined #shogun | 15:45 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 15:51 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has joined #shogun | 15:51 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 16:11 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has joined #shogun | 16:12 | |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun | 16:15 | |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has quit [Quit: Page closed] | 17:30 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 17:49 | |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has joined #shogun | 17:51 | |
-!- SCV__ [~FSCV@187.210.54.165] has joined #shogun | 18:24 | |
-!- FSCV_ [~FSCV@204.45.132.11] has quit [Ping timeout: 245 seconds] | 18:24 | |
blackburn | heiko: have you ever played fallout? | 19:25 |
blackburn | :) | 19:25 |
heiko | blackburn, why you ask? | 19:25 |
heiko | yes | 19:26 |
blackburn | heiko: I stumbled upon a photo and created this https://dl.dropbox.com/u/10139213/share/fallout.jpg | 19:26 |
blackburn | :D | 19:26 |
heiko | haha | 19:26 |
blackburn | pure russia | 19:26 |
heiko | :) | 19:26 |
blackburn | magnitogorsk | 19:26 |
heiko | well, I have played it :) | 19:27 |
blackburn | heiko: that one with a cigarette looks like a rad-monster! | 19:27 |
blackburn | with camouflage :D | 19:27 |
heiko | ha! | 19:27 |
heiko | nice one :) | 19:27 |
blackburn | heiko: http://ic.pics.livejournal.com/drugoi/484155/7591058/original.jpg that's what it was before my GIMP | 19:27 |
heiko | man thats even worse | 19:28 |
blackburn | heiko: lol I've accidentally hidden factories | 19:28 |
heiko | yes | 19:28 |
heiko | so you made the image less-falloutish | 19:29 |
blackburn | hahah | 19:29 |
blackburn | lol | 19:29 |
blackburn | heiko: I should use that pic to convince sonney2k that things are not the same in different countries | 19:31 |
blackburn | :D | 19:31 |
heiko | why doesnt he believe this? :) | 19:31 |
blackburn | heiko: no idea :) but he even wasn't even surprised when I told him that a man was killed with a champagne bottle in police | 19:32 |
blackburn | double even hah | 19:32 |
heiko | Russian is a wild country I guess | 19:33 |
blackburn | heiko: sure | 19:36 |
blackburn | you will see once you see me | 19:36 |
heiko | haha :) | 19:37 |
heiko | so many scars? | 19:37 |
blackburn | heiko: hmm don't have any :( | 19:38 |
heiko | mmh, well then I got more than you :) | 19:38 |
blackburn | heiko: I was asked about combining linear and gaussian kernels | 19:38 |
heiko | blackburn, and? | 19:38 |
blackburn | linear is like in [-1,1] | 19:38 |
blackburn | and gaussian is in [0,1] | 19:38 |
blackburn | anything to do with it? | 19:39 |
heiko | what means "in"? | 19:39 |
blackburn | in range | 19:39 |
heiko | I dont really get what you mean | 19:42 |
heiko | the range of the kernel function? | 19:43 |
blackburn | heiko: they have different value domains yes | 19:43 |
blackburn | :) | 19:43 |
heiko | and now they have to be combined? | 19:43 |
blackburn | yeah - is it ok? | 19:43 |
heiko | as long as the linear kernel matrix is psd, this is fine | 19:43 |
heiko | the linear one wont be negative when the argument is equal right? | 19:44 |
heiko | since that would destroy it | 19:44 |
blackburn | yeah psd is not a problem | 19:44 |
blackburn | linear combination is still mercer kernel I know | 19:44 |
blackburn | but just these different domains - I never though about that | 19:45 |
heiko | no that is fine | 19:46 |
heiko | kernel always goes to real numbers | 19:46 |
heiko | since it is an inner product | 19:46 |
blackburn | heiko: no I am not talking about C :) | 19:47 |
heiko | what? | 19:48 |
blackburn | heiko: complex numbers | 19:49 |
blackburn | I mean :) | 19:49 |
heiko | why should this negative range be a problem? It is allowed since a kernel is a map X x X => R where X is some set | 19:49 |
heiko | so what problem should be there? | 19:49 |
blackburn | hah no problem I just find that interesting | 19:49 |
-!- heiko [~heiko@nat-164-84.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 20:16 | |
-!- tomv_ [~tomv@p4FD61DE8.dip.t-dialin.net] has joined #shogun | 21:07 | |
tomv_ | hi. so with blackburn's help I got shogun for python running. but now I am wondering: the FITCInferenceMethod seems to only work when the number of latent features is equal to the number of features. | 21:08 |
blackburn | hi tomv_ | 21:09 |
tomv_ | This is surprising to me, having looked at gpml and also read rasmussen and williams' book... | 21:09 |
blackburn | tomv_: that could be hard as I never managed to get throught rasmussens' book | 21:10 |
tomv_ | well, really the sparse stuff is much better in the overview article IMHO | 21:10 |
blackburn | Regular and Latent Features do not match in dimensionality | 21:11 |
blackburn | you get that, right? | 21:11 |
tomv_ | yes. | 21:12 |
tomv_ | well, it is unfair to call it a review article. I had been looking at "A Unifying View of Sparse Approximate Gaussian Process" | 21:12 |
tomv_ | or looking at the Large Scale Regression section on http://www.gaussianprocess.org/gpml/code/matlab/doc/ | 21:13 |
blackburn | tomv_: do you have some reproducing code? | 21:13 |
tomv_ | gpml does. | 21:13 |
blackburn | argh heiko has left the room already | 21:14 |
tomv_ | if you look at the picture, the circles are the latent features. unless latent features is something other than inducing inputs. | 21:14 |
blackburn | tomv_: did yuo check that example on fitc we have in examples/undocumented/libshogun? | 21:15 |
tomv_ | actually I used the one in examples/documented/libshogun and tried to port it to python to give a small token of appreciation for your help. | 21:16 |
blackburn | tomv_: so black circles are kind of latent variables? | 21:17 |
tomv_ | but then I stumbled over my idea of how FITC works being literally incompatible to shogun's(?) | 21:17 |
tomv_ | python: /usr/include/eigen3/Eigen/src/Core/DenseCoeffsBase.h:352: Scalar &Eigen::DenseCoeffsBase<Eigen::Map<Eigen::Matrix<double, -1, -1, 0, -1, -1>, 0, Eigen::Stride<0, 0> >, 1>::operator()(Index, Index) [Derived = Eigen::Map<Eigen::Matrix<double, -1, -1, 0, -1, -1>, 0, Eigen::Stride<0, 0> >, Level = 1]: Assertion `row >= 0 && row < rows() && col >= 0 && col < cols()' failed. | 21:17 |
tomv_ | Aborted | 21:17 |
blackburn | uh | 21:17 |
tomv_ | is what I get. | 21:17 |
tomv_ | yes, that is my understanding | 21:17 |
blackburn | I have no understanding so far as I am stupid code monkey that knows only dimension reduction :D | 21:18 |
blackburn | could you share that failing code? | 21:18 |
tomv_ | sure. | 21:18 |
blackburn | tomv_: are you subscribed to the mailing list? | 21:21 |
tomv_ | http://vman.de/shogun/regression_gaussian_process_fitc_demo.py | 21:23 |
tomv_ | no | 21:23 |
blackburn | I am asking because there is a guy who can help you with GPs I think | 21:23 |
blackburn | heiko strathmann that is - I see you can even talk in person in your common language with more understanding :) | 21:24 |
blackburn | man I like your homepage :) | 21:25 |
blackburn | ok let me check example now | 21:25 |
blackburn | tomv_: I get another error message, not eigen3 one | 21:29 |
tomv_ | fascinating. | 21:29 |
tomv_ | what do you get? | 21:29 |
blackburn | something about parameters | 21:29 |
blackburn | I'll try to work it out | 21:30 |
tomv_ | have you compiled without nlopt | 21:30 |
tomv_ | ? | 21:30 |
blackburn | ahh right | 21:30 |
tomv_ | ha. | 21:30 |
blackburn | okay let me compile with nlopt then | 21:30 |
blackburn | hmmm wait | 21:30 |
blackburn | it is with nlopt already | 21:30 |
blackburn | surprise | 21:30 |
blackburn | anyway recompiling | 21:31 |
blackburn | I get segfault actually | 21:31 |
tomv_ | that should not happen in python, either. though the abort is really bad, too. | 21:32 |
tomv_ | maybe you have optimised the assetions away? | 21:32 |
tomv_ | don't know if that can happen with nlopt | 21:32 |
tomv_ | I am using the stock debian/sid one. | 21:32 |
blackburn | tomv_: segfault happens with python - why not? shogun is C++ | 21:34 |
-!- n4nd0 [53b32c87@gateway/web/freenode/ip.83.179.44.135] has joined #shogun | 21:36 | |
n4nd0 | blackburn: really nice form | 21:46 |
blackburn | n4nd0: thanks to goolge :) | 21:47 |
tomv_ | blackburn: well, segfault in python means the lib has a bug. | 21:49 |
blackburn | tomv_: exactly | 21:50 |
n4nd0 | tomv_, blackburn where do we have a segfault? | 22:01 |
blackburn | n4nd0: in parameters stuff | 22:01 |
n4nd0 | mmm I see | 22:01 |
tomv_ | actually, I don't have one, I get assertion errors (which, I would imagine, would lead to a segfault if the assertion is skipped due to optimization options in nlopt) | 22:02 |
n4nd0 | tomv_: mmm if you get an assertion error then execution *must* stop | 22:15 |
n4nd0 | ahhh ok I see what you mean | 22:16 |
tomv_ | hm, why? | 22:16 |
n4nd0 | well because the assertion works like that | 22:16 |
n4nd0 | assert(boolean expression) | 22:16 |
n4nd0 | if the boolean expression is false, then stop | 22:16 |
n4nd0 | tomv_: agree? | 22:16 |
tomv_ | well, to be honest, I might think that exiting the python interpreter because a function call ran into an assertion error is excessive. | 22:17 |
tomv_ | it's assert, not falloverifitsnot. | 22:17 |
n4nd0 | tomv_: mmm | 22:17 |
n4nd0 | but an assert is used to stop if the expression given is false right? | 22:18 |
tomv_ | yes. | 22:18 |
n4nd0 | I mean say I want to compute the distances between two sets of features | 22:18 |
n4nd0 | if one of them is empty | 22:18 |
n4nd0 | then it makes sense to stop rightaway | 22:18 |
n4nd0 | because it doesn't make sense to compute the distance between some features and nothing | 22:19 |
tomv_ | well, actually, the c++ spec says that the program should abort. | 22:20 |
tomv_ | so yes. | 22:20 |
tomv_ | not "spec" , reference | 22:20 |
n4nd0 | tomv_: ok anyway, all this was almost philosophic :) | 22:20 |
n4nd0 | what do you have in your code a failed assertion or a seg fault? | 22:21 |
tomv_ | and it says that compiling with NDEBUG the assertions are stricken. | 22:21 |
tomv_ | I have a failed assertion, blackburn has the segfault. | 22:21 |
tomv_ | http://vman.de/shogun/regression_gaussian_process_fitc_demo.py | 22:21 |
tomv_ | running the above. | 22:21 |
n4nd0 | ok | 22:23 |
n4nd0 | I cannot test it right now | 22:24 |
n4nd0 | but if you have a failed assertion, I'd suggest to check the inputs you are giving there | 22:24 |
tomv_ | I am afraid sleeping might be the first order of business for me. good night and thanks for your interest. | 22:26 |
-!- tomv_ [~tomv@p4FD61DE8.dip.t-dialin.net] has quit [Quit: Ciao] | 22:26 | |
n4nd0 | night :) | 22:27 |
-!- heiko1 [~heiko@b0f96dbb.bb.sky.com] has joined #shogun | 22:35 | |
blackburn | heiko1: ! | 22:35 |
blackburn | we missed you | 22:35 |
heiko1 | I hope so :) | 22:35 |
blackburn | heiko1: there was a guy who complained about GPs | 22:35 |
heiko1 | complained? | 22:35 |
blackburn | heiko1: yes it doesn't work | 22:36 |
blackburn | :D | 22:36 |
heiko1 | blackburn, I know | 22:36 |
blackburn | heiko1: do you know about FITC? | 22:36 |
heiko1 | Jacob implemented that riught? | 22:37 |
blackburn | exactly | 22:37 |
heiko1 | I couldnt even make regression work when I tried it :) | 22:37 |
blackburn | but you said you are going to be a GP master | 22:37 |
heiko1 | Yeah, I want to take this over | 22:37 |
blackburn | I have a feeling you didn't so far | 22:37 |
blackburn | :D | 22:37 |
heiko1 | worked with GP over the last months | 22:37 |
blackburn | okayface | 22:37 |
heiko1 | well its a lot of work | 22:38 |
heiko1 | quite a big framwork | 22:38 |
heiko1 | and not so much time here :) | 22:38 |
heiko1 | I even reserached new methods that I want to implement :) | 22:38 |
blackburn | I am confused with a lot of things in shogun | 22:38 |
heiko1 | but gotta do something else now | 22:38 |
heiko1 | so see you in a bit | 22:38 |
blackburn | heiko1: okay | 22:38 |
n4nd0 | blackburn: so how many days do you think we will be shoguning in Berlin? | 23:01 |
blackburn | n4nd0: I expect only tow | 23:02 |
blackburn | :) | 23:02 |
blackburn | two | 23:02 |
n4nd0 | it sounds fine | 23:02 |
n4nd0 | the plan is to see if we can be at TU the second day right? | 23:02 |
n4nd0 | and what is this lookalike contest?! hehe | 23:03 |
blackburn | yeah | 23:03 |
blackburn | just joke (*I hope*) | 23:03 |
n4nd0 | are we going to look for people that look like you? hehe | 23:03 |
n4nd0 | I think the competition thing for the audience is a very nice idea | 23:04 |
n4nd0 | btw | 23:06 |
n4nd0 | nvm | 23:06 |
blackburn | nvm? | 23:13 |
blackburn | :) | 23:13 |
n4nd0 | hehe yeah was going to ask sth but I answered myself instead :D | 23:19 |
@sonney2k | blackburn, the issue tomv_ is having is the one I am debugging... | 23:58 |
@sonney2k | no success yet though | 23:58 |
* sonney2k is happy to be online again | 23:58 | |
@sonney2k | bricked the router and hat to short circuit some pins on the flash chip | 23:59 |
--- Log closed Tue Jan 29 00:00:38 2013 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!