--- Log opened Sun Jul 10 00:00:59 2011 | ||
* blackburn got insight of various abracadabras DGEMM DPOTRI | 00:02 | |
CIA-32 | shogun: Sergey Lisitsyn master * r67826b5 / (src/libshogun/lib/arpack.cpp src/libshogun/lib/arpack.h): Added Cholesky factorization for Arpack wrapper - http://bit.ly/oDnBVv | 00:35 |
---|---|---|
CIA-32 | shogun: Sergey Lisitsyn master * r46aed74 / (2 files): Modified preprocessors using arpack for new parameter - http://bit.ly/qiTfZR | 00:35 |
blackburn | SHHHHH | 00:35 |
CIA-32 | shogun: Sergey Lisitsyn master * r419387b / (2 files in 2 dirs): Removed junk - http://bit.ly/p3mtEP | 00:37 |
blackburn | unfortunately lle produces indefinite symmetrical matrix :( | 00:38 |
-!- f-x [~user@117.192.197.89] has quit [Remote host closed the connection] | 01:21 | |
-!- blackburn [~blackburn@188.122.238.99] has quit [Quit: Leaving.] | 01:56 | |
-!- in3xes_ [~in3xes@180.149.49.227] has joined #shogun | 05:25 | |
-!- in3xes [~in3xes@210.212.58.111] has quit [Ping timeout: 240 seconds] | 05:28 | |
-!- in3xes_ is now known as in3xes | 05:34 | |
-!- [1]warpy [~warpy@bzq-79-179-0-217.red.bezeqint.net] has joined #shogun | 06:50 | |
[1]warpy | hey guys anyone alive | 06:56 |
bettyboo | roger | 06:56 |
[1]warpy | i was wondering how to download 0.11 | 06:56 |
-!- [1]warpy [~warpy@bzq-79-179-0-217.red.bezeqint.net] has quit [Quit: HydraIRC -> http://www.hydrairc.com <- In tests, 0x09 out of 0x0A l33t h4x0rz prefer it :)] | 07:34 | |
-!- in3xes_ [~in3xes@180.149.49.230] has joined #shogun | 09:02 | |
-!- in3xes_ [~in3xes@180.149.49.230] has quit [Remote host closed the connection] | 09:29 | |
-!- shayan_ [~shayan@180.149.49.225] has joined #shogun | 12:39 | |
-!- shayan_ is now known as shayan | 13:02 | |
-!- shayan [~shayan@180.149.49.225] has quit [Quit: Leaving] | 13:25 | |
-!- f-x [~user@117.192.203.242] has joined #shogun | 14:32 | |
-!- blackburn [~blackburn@188.122.238.99] has joined #shogun | 18:32 | |
-!- f-x [~user@117.192.203.242] has quit [Remote host closed the connection] | 18:33 | |
blackburn | got kindle dx! now can read papers without any hurt ;) | 18:42 |
-!- blackburn [~blackburn@188.122.238.99] has quit [Read error: No route to host] | 19:32 | |
-!- blackburn [~blackburn@188.122.238.99] has joined #shogun | 19:32 | |
-!- f-x [~user@117.192.203.242] has joined #shogun | 20:35 | |
-!- f-x [~user@117.192.203.242] has left #shogun [] | 20:35 | |
-!- f-x [~user@117.192.203.242] has joined #shogun | 20:35 | |
@sonney2k | blackburn, without any hurt? | 22:22 |
blackburn | sonney2k: yeah reading using laptop suckz, aint it? | 22:22 |
@sonney2k | blackburn, yeah but is this A4 size / I mean don't you need to zoom the pdf to read anything? | 22:23 |
blackburn | sonney2k: most papers have large margins - so easy readable | 22:24 |
CIA-32 | shogun: Baozeng Ding master * rb0c3848 / (4 files): add distribution java examples - http://bit.ly/oqMg27 | 22:24 |
CIA-32 | shogun: Soeren Sonnenburg master * r268bccc / (4 files): | 22:24 |
CIA-32 | shogun: Merge pull request #185 from sploving/master | 22:24 |
CIA-32 | shogun: add remaining distribution java examples - http://bit.ly/mY9Cbp | 22:24 |
@sonney2k | blackburn, ok... | 22:24 |
blackburn | e.g. elements of statistical learning looks pretty without zoom | 22:24 |
blackburn | as is | 22:24 |
@sonney2k | hmmhh... looks like things are progressing | 22:24 |
blackburn | which things? | 22:24 |
@sonney2k | I should probably start test shogun compilation on osx / cygwin | 22:25 |
blackburn | I'm now trying to get causes of LLE strange results sometimes | 22:25 |
-!- blackburn [~blackburn@188.122.238.99] has left #shogun [] | 22:28 | |
-!- blackburn [~blackburn@188.122.238.99] has joined #shogun | 22:28 | |
@sonney2k | blackburn, that is actually also a good candidate for shogun http://mloss.org/software/view/320/ | 22:34 |
@sonney2k | all the possible multiclass svms... | 22:34 |
blackburn | sonney2k: where to get so much consistent developers? ;) | 22:34 |
@sonney2k | consistent? | 22:35 |
@sonney2k | did you mean where to get enough developers to support all the many features? | 22:35 |
blackburn | ehh sth like that | 22:36 |
blackburn | well where to get any developers? :) | 22:37 |
@sonney2k | blackburn, well year over year there must be another GSoC :) | 22:37 |
blackburn | yeah I hope to get into it ;) | 22:38 |
* sonney2k looks at the mc svm code stuff | 22:46 | |
@sonney2k | interesting | 22:46 |
@sonney2k | they have sse4.1 tuned dot products in there | 22:46 |
blackburn | hoho | 22:46 |
@sonney2k | blackburn, ^ maybe we should benchmark and borrow their code :) | 22:47 |
blackburn | well if it isn't a task for a 1.0 - I could do it some day | 22:47 |
blackburn | I should be 'vacant' for ~10 hours per week for some shoguning after gsoc | 22:48 |
blackburn | sonney2k: do we have some up-to-date multiclass svm now? | 22:50 |
@sonney2k | blackburn, we have GMNPSVM | 22:50 |
blackburn | it tells me almost nothing :) | 22:50 |
@sonney2k | and some one-vs-one via LibSVM | 22:50 |
@sonney2k | but that's it | 22:51 |
blackburn | sonney2k: is there any not one-vs-one for SVMs? | 22:51 |
@sonney2k | the one in the package were true MC SVMs | 22:51 |
@sonney2k | GMNP* | 22:51 |
blackburn | how it works? | 22:51 |
@sonney2k | and one-vs-rest one has to do manually currently | 22:51 |
blackburn | ah | 22:51 |
@sonney2k | blackburn, it is a true multiclass svm | 22:51 |
blackburn | I wonder how it can be formulated | 22:52 |
@sonney2k | easy: you want f(x,y) for the true class y to be greater than f(x,y') for y' not the true class | 22:52 |
@sonney2k | and f is a linear function again - but rest is the same as svm | 22:53 |
@sonney2k | optimization problem is harder though | 22:53 |
blackburn | hmm | 22:53 |
@sonney2k | blackburn, I mean in SVM you usually just say points should be on the right side of the hyperplane | 22:57 |
blackburn | yeap | 22:57 |
@sonney2k | now you say output of the function for the true label should be greater than output of the function for all wrong labels plus some margin | 22:57 |
@sonney2k | so you have many more constraints | 22:57 |
blackburn | ah seems I understood | 22:58 |
blackburn | :( got dataset where LLE produces different results for different solvers | 23:00 |
blackburn | debug session #N | 23:01 |
blackburn | not stable problem I guess | 23:02 |
blackburn | sonney2k: I've finally changed everything to vim :) because of you! | 23:06 |
@sonney2k | I am sorry | 23:06 |
blackburn | :D | 23:06 |
-!- f-x [~user@117.192.203.242] has quit [Remote host closed the connection] | 23:06 | |
blackburn | sonney2k: well it is cool, I understand, I guess you (with some experience and skillz) use it almost magically | 23:07 |
@sonney2k | I don't even use all the features. | 23:09 |
@sonney2k | but I got very used to it now... | 23:09 |
@sonney2k | do you know whether eclipse has a vim plugin? | 23:09 |
@sonney2k | http://eclim.org/ | 23:10 |
blackburn | sonney2k: yes, eclipse have one, I know | 23:10 |
@sonney2k | very good | 23:10 |
@sonney2k | if I ever have to use eclipse I can at least feel home :) | 23:11 |
bettyboo | :) | 23:11 |
blackburn | for java I consider ide ;) | 23:11 |
blackburn | sonney2k: is it *normal* in ML that some methods can fail if say k is bad? | 23:23 |
@sonney2k | blackburn, welcome to numerics | 23:23 |
@sonney2k | sure | 23:24 |
blackburn | e.g. LLE sometimes produces examples lying on diagonal line | 23:24 |
blackburn | not that beauty I showed you yesterday | 23:24 |
@sonney2k | some methods are very sensitive | 23:24 |
@sonney2k | some very robust | 23:24 |
blackburn | I'll check it for sure - but looks like it is not robust | 23:24 |
blackburn | sparse matrix eigenproblem doesn't look robust for me, am I right? | 23:25 |
blackburn | voices in my head tells me it is not, sonney2k, do you have any? | 23:26 |
@sonney2k | blackburn, I am not an expert in this subject but I would guess so :) | 23:28 |
blackburn | me too, but voices.. :D | 23:28 |
blackburn | scikit-learn started as GSoC project?! | 23:33 |
@sonney2k | blackburn, I am currently checking java examples | 23:40 |
@sonney2k | Isomap preprocessor = new Isomap() | 23:40 |
@sonney2k | gives an error | 23:40 |
@sonney2k | does it not exist or has it a different name? | 23:40 |
blackburn | sonney2k: isomap is abstract class | 23:40 |
blackburn | ClassicIsomap or LandmarkIsomap | 23:40 |
blackburn | better both :D | 23:41 |
@sonney2k | yes that works | 23:41 |
@sonney2k | blackburn, btw, did you touch / polish kernelpca / pca? | 23:41 |
blackburn | not yet | 23:41 |
blackburn | arpack issues.. | 23:42 |
@sonney2k | I see | 23:42 |
@sonney2k | blackburn, good luck :) | 23:42 |
blackburn | now writing to jake vanderplas, he have implemented lle in scikits | 23:42 |
blackburn | I really warned with some unstability | 23:43 |
CIA-32 | shogun: Soeren Sonnenburg master * ree98378 / (3 files in 2 dirs): fix java examples - http://bit.ly/qgs5tG | 23:44 |
CIA-32 | shogun: Soeren Sonnenburg Siddharthk-master * r8ed3110 / (4 files in 3 dirs): Merge branch 'master' of https://github.com/Siddharthk/shogun into Siddharthk-master - http://bit.ly/qfWaHb | 23:44 |
--- Log closed Mon Jul 11 00:00:05 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!