--- Log opened Wed Nov 14 00:00:17 2012 | ||
@sonney2k | ptizoom, excellent work as usual... | 00:09 |
---|---|---|
blackburn | la la | 00:09 |
-shogungit:#shogun- [shogun] sonney2k pushed 3 new commits to master: https://github.com/shogun-toolbox/shogun/compare/742b70c55a8c...2f3bb6289502 | 00:10 | |
-shogungit:#shogun- shogun/master c3dd767 ptizoom: perl_modular: fixed matrix SG to PDL conversion.... | 00:10 | |
-shogungit:#shogun- shogun/master f566eba ptizoom: tests/perl_modular: increase max_train_time of linear classifiers | 00:10 | |
-shogungit:#shogun- shogun/master 2f3bb62 Soeren Sonnenburg: Merge pull request #822 from ptizoom/perl_121113... | 00:10 | |
-!- blackburn [~blackburn@46.20.65.191] has quit [Quit: Leaving.] | 00:56 | |
-!- ptizoom [~christian@85.210.94.175] has quit [Quit: Ex-Chat] | 03:13 | |
-!- ptizoom [~christian@85.210.94.175] has joined #shogun | 03:14 | |
-!- zxtx [~zv@cpe-75-83-151-252.socal.res.rr.com] has joined #shogun | 07:48 | |
ptizoom | blackburn:yet another exemple of backward propagation training...reward from the master, jealousy from the others.... | 08:55 |
sonne|work | work == reward, more work == more reward | 10:44 |
sonne|work | wiking: I will ask today. | 10:44 |
sonne|work | this situation is not really helpful | 10:45 |
wiking | sonne|work: yo | 10:53 |
wiking | sonne|work: yeah but i think they have their 'good reason' for it | 10:53 |
wiking | so by asking we are not going to go any further | 10:53 |
wiking | i guess the latest we'll know something about it on the 3rd of december :D | 10:53 |
wiking | when we realise that we are not there or when we are ther ;) | 10:54 |
wiking | *there | 10:54 |
sonne|work | hehe | 10:56 |
wiking | sonne|work: you should make a shogun-foundation ;) | 10:57 |
wiking | sonne|work: where people can donate :> | 10:57 |
sonne|work | wiking: well I asked upfront and carol said that they know that the schedule is extremely tight so they want to do the decision w/i a week | 10:57 |
wiking | they can even make donation for a given feature ;) | 10:58 |
wiking | that'd be cool | 10:58 |
wiking | :D | 10:58 |
sonne|work | haha | 10:58 |
sonne|work | carol is on vacations | 10:58 |
wiking | hehehe | 10:58 |
wiking | mm isn't thanksgiving is about to happen soon? | 10:58 |
sonne|work | returning just before the meeting :) | 10:59 |
wiking | oh yeah 22nd of nov | 10:59 |
wiking | so maybe she went on hunting some turkey ;P | 11:00 |
wiking | sonne|work: here? | 13:27 |
sonne|work | yes? | 13:27 |
wiking | sonne|work: any suggestion which machine would be adecvate to solve this dual problem: | 13:28 |
wiking | i hate pdf | 13:28 |
wiking | sorry gimme a sec to formulate the problem | 13:28 |
wiking | sonne|work: http://test.izyba.com/latex?&space;D%28\alpha_*,\beta_*%29=max_{\alpha,%20\beta}%20\sum_i%20\alpha_i%20+%20\sum_j\sum_h%20\beta_{j,h}%20-%20\frac{1}{2}%20||\sum_{\alpha}%20\alpha%20\Phi%28x_i,%20h_i%29%20-%20\sum_j%20\sum_h%20\beta_{j,h}%20\Phi%28x_j,%20h%29||^2%20\\%20\mattext{s.t.%200%20%3C=%20\alpha_i%20%3C=%20C_i,%20\forall%20i;%200%20%3C=%20\beta_{j,h}%20%3C=%20C_2,%20\forall%20j,%20\forall%20\in%20H} | 13:33 |
wiking | shiiit | 13:33 |
wiking | sonne|work: this is the problem http://pastebin.com/X2t8bEAd | 13:34 |
wiking | and just copy-paste it there: http://www.codecogs.com/latex/eqneditor.php | 13:34 |
sonne|work | http://mathbin.net/113694 | 13:36 |
wiking | yeeey | 13:39 |
wiking | that :D | 13:39 |
sonne|work | it looks very SVMish | 13:42 |
wiking | yeah it's a dual form | 13:42 |
sonne|work | yeah but you also only have box constraints on alpha / beta | 13:50 |
sonne|work | you should make the quadratic term explicit btw | 13:51 |
sonne|work | you could certainly use ocas to solve that | 13:51 |
wiking | ocas? | 13:53 |
sonne|work | well libqp | 13:53 |
wiking | aaah | 13:53 |
sonne|work | ocas assumes fixed Cs | 13:53 |
wiking | mmm | 13:53 |
wiking | i'm just checking libqp's formulation | 13:53 |
sonne|work | I mean you could introduce gamma = [alpha beta] and reformulate the problem to use only that | 13:54 |
sonne|work | then you have just one variable to optimize over | 13:54 |
sonne|work | different box constraints for different parts | 13:54 |
sonne|work | and some linear term | 13:54 |
sonne|work | and quadratic term | 13:54 |
sonne|work | http://cmp.felk.cvut.cz/~xfrancv/libqp/html/ | 13:55 |
wiking | yeah | 13:55 |
sonne|work | you have no bias term so no equality constraint even | 13:55 |
sonne|work | so the problem is more simple but I don't know if libqp has something internally that works w/o one | 13:56 |
wiking | but this is minimizing the problem :S | 13:57 |
wiking | i mean qp | 13:58 |
wiking | *libqp | 13:58 |
sonne|work | then mimimize -qp :D | 13:59 |
wiking | mmmm | 14:00 |
wiking | i hope S_{equ} can be a zeroset | 14:01 |
sonne|work | no not allowed | 14:04 |
sonne|work | ask vojtech | 14:04 |
wiking | :S | 14:05 |
sonne|work | I am pretty sure one of the sub-solvers for solving the problem w/ bias constraint needs a solver w/o one - so that one should be in libqp too | 14:06 |
sonne|work | btw I guess you tested this with mosek already? | 14:06 |
wiking | no not at all | 14:07 |
sonne|work | why not? | 14:08 |
wiking | i've just seen this paper | 14:12 |
wiking | and it would be great to have a solver that doesn't depend on mosek :) | 14:12 |
sonne|work | sure but first build a reference then do the real impl. | 14:13 |
wiking | mmm i guess mosek has a solver for this | 14:15 |
sonne|work | sure it is a QP | 14:19 |
sonne|work | but you still need to create that gamma | 14:19 |
wiking | ah yeah | 14:22 |
wiking | the question still the same | 14:22 |
wiking | sonne|work: https://gist.github.com/4072038 | 14:23 |
wiking | line 197 | 14:23 |
wiking | i want to commit this solver tday | 14:24 |
wiking | so i need a solution ;) | 14:24 |
sonne|work | its only in a .cpp so I don't care | 14:35 |
wiking | sonne|work: ah sure? | 14:38 |
wiking | ok then here goes the PR | 14:38 |
-shogungit:#shogun- [shogun] vigsterkr opened pull request #823: Add SO-SVM solver based on Chun-Nam Yu's implementation (master...so) https://github.com/shogun-toolbox/shogun/pull/823 | 14:58 | |
sonne|work | wiking: did you ifdef with HAVE_MOSEK ? | 14:59 |
sonne|work | wiking: and could you do w/o vector in the .h? | 14:59 |
wiking | i've done ifdef for mosek | 15:00 |
sonne|work | apart from that ok | 15:00 |
wiking | since this solver could use SVMlight | 15:00 |
sonne|work | cannot see it | 15:00 |
wiking | it's only within cpp at a given place | 15:00 |
wiking | https://github.com/shogun-toolbox/shogun/pull/823/files#L0R329 | 15:00 |
wiking | ah ok i'll remove the std::vector header includ ethere... | 15:01 |
wiking | but we need to update svmlight somehow for that | 15:04 |
wiking | :) | 15:04 |
wiking | sonne|work: ok i cannot move the include ;) | 15:07 |
wiking | sonne|work: https://github.com/shogun-toolbox/shogun/pull/823/files#L1R137 | 15:08 |
wiking | as the u can see the resize_cleanup function args contain std::vector | 15:08 |
-shogungit:#shogun- [shogun] ptizoom opened pull request #824: perl_modular: fixed string_from_pdl<char> conversion. (master...perl_121114) https://github.com/shogun-toolbox/shogun/pull/824 | 16:19 | |
sonne|work | wiking: couldn't you use SGVector for that or dynarray? | 16:23 |
-shogungit:#shogun- [shogun] sonney2k pushed 2 new commits to master: https://github.com/shogun-toolbox/shogun/compare/2f3bb6289502...0cdc7eb45a44 | 16:29 | |
-shogungit:#shogun- shogun/master d2765ab ptizoom: perl_modular: fixed string_from_pdl<char> conversion.... | 16:29 | |
-shogungit:#shogun- shogun/master 0cdc7eb Soeren Sonnenburg: Merge pull request #824 from ptizoom/perl_121114... | 16:29 | |
wiking | sonne|work: i'd need to define the template for SGVector<SGSpsarseVector...> | 16:36 |
wiking | but i need to send a fix anyways for that PR | 16:37 |
wiking | i'm just checking whether it compiles now | 16:37 |
sonne|work | wiking: btw do you understand why std::vector works even with SGVector objects? | 16:39 |
sonne|work | I mean it is on the stack | 16:39 |
sonne|work | or is it using new[] ? | 16:39 |
sonne|work | wiking: anyway you could use DynArray | 16:40 |
sonne|work | but you need these sh... inplace constructors | 16:40 |
wiking | sonne|work: it's on the heap | 16:40 |
wiking | there... let's see if travis can compile it now | 16:43 |
wiking | ok at least now it compiles on travis | 16:48 |
ptizoom | sonne*:red concerns on formatting...I cannot promise a good style, but suggest to indent at the end of development and doc....but will try anyway! | 17:01 |
ptizoom | sonne*:there are still many big problems in perl.... | 17:02 |
ptizoom | sonne*:sparse stuff is not tested at all! inf and nan values handling. | 17:03 |
ptizoom | sonne*: also, string_to_pdl does not work as expected...I was looking at that. | 17:04 |
-!- blackburn [~blackburn@37.61.179.13] has joined #shogun | 17:48 | |
blackburn | wiking: oppan gangnam style solver? | 17:48 |
blackburn | sonne|work: yes std vector uses constructors | 18:26 |
blackburn | so no problem with sgvector | 18:26 |
blackburn | std is safe you should believe :) | 18:27 |
wiking | blackburn: :DDDDDDDDDDDDDDDDD | 19:09 |
@sonney2k | wiking, I just learned - tonight (timeout of your choice) we will get some notification email | 22:02 |
@sonney2k | so so after lunch tomorrow CET | 22:03 |
@sonney2k | s/timeout/timezone | 22:03 |
wiking | :> | 22:18 |
wiking | i guess that means we are not in :) | 22:18 |
wiking | otherwise they would have sent that yeah you are in but we'll announce it officially tonight :) | 22:19 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 22:37 | |
* wiking having the flow with boost | 23:10 | |
blackburn | wiking: boost is illegal in germany | 23:12 |
wiking | idontcare | 23:16 |
wiking | i'm working on my shit | 23:16 |
wiking | :) | 23:16 |
blackburn | :D | 23:16 |
blackburn | we are all working on our own shit | 23:16 |
wiking | somebody figure out wha tto do with that std::vector so that it can be merged | 23:17 |
blackburn | wiking: I have nothing against it in .cpp | 23:18 |
wiking | its in .h | 23:18 |
wiking | as for a func args | 23:18 |
blackburn | ah | 23:18 |
blackburn | wiking: may be some vodka? | 23:19 |
wiking | :> | 23:19 |
blackburn | hard to speed up brain activity w/o | 23:19 |
wiking | :> | 23:23 |
wiking | yes yes | 23:23 |
wiking | hahaha i think the most i like about c++11 is the auto variable :) | 23:33 |
blackburn | why? | 23:33 |
wiking | no more fucking | 23:33 |
wiking | std::map<something long, somethingevenlonger>::const_iterator | 23:34 |
blackburn | typedef | 23:34 |
wiking | lolface | 23:34 |
blackburn | you should not use such types in your code I think | 23:34 |
wiking | i'm not creating industry standard code | 23:34 |
blackburn | I use typedefs for that | 23:34 |
wiking | just trying to test something as fast as possible | 23:34 |
wiking | and usually at that moment | 23:34 |
blackburn | I'd rather have ParameterMap than std::unordered_map<fdsfsdfdsf,fsdfsdfs> | 23:34 |
wiking | my code changes like everyfuckingtime | 23:35 |
--- Log closed Thu Nov 15 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!