--- Log opened Tue Jun 03 00:00:16 2014 | ||
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Ping timeout: 240 seconds] | 00:19 | |
shogun-buildbot_ | build #327 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/327 blamelist: abinashpanda <abinash.panda.ece10@itbhu.ac.in> | 00:21 |
---|---|---|
shogun-buildbot_ | build #2989 of deb1 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/deb1%20-%20libshogun/builds/2989 | 00:23 |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has joined #shogun | 00:25 | |
-!- thor|home [~thoralf@91-65-141-142-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 00:28 | |
-!- khalednasr [~k.nasr92@41.69.214.54] has joined #shogun | 00:36 | |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has joined #shogun | 00:36 | |
shogun-buildbot_ | build #657 of precise - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/657 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com> | 00:44 |
-!- khalednasr [~k.nasr92@41.69.214.54] has quit [Quit: Leaving] | 01:41 | |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Ping timeout: 260 seconds] | 01:51 | |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 01:52 | |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has joined #shogun | 02:04 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 02:27 | |
shogun-buildbot_ | build #18 of nightly_fedora is complete: Failure [failed git] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_fedora/builds/18 blamelist: abinashpanda <abinash.panda.ece10@itbhu.ac.in>, Fernando Iglesias <fernando.iglesiasg@gmail.com>, tklein23 <tklein23@users.noreply.github.com>, Parijat Mazumdar <mazumdarparijat@gmail.com> | 03:00 |
shogun-buildbot_ | build #328 of debian wheezy - memcheck is complete: Failure [failed memory check] Build details are at http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/328 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com> | 03:14 |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Ping timeout: 252 seconds] | 03:44 | |
-!- kislay [~Abhijeet@103.240.205.197] has quit [Quit: Leaving] | 04:01 | |
-!- zxtx [~zv@98.223.196.32] has joined #shogun | 04:38 | |
-!- kislay [~Abhijeet@42.104.13.186] has joined #shogun | 05:27 | |
-!- kislay [~Abhijeet@42.104.13.186] has quit [Client Quit] | 05:27 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 07:16 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 07:16 | |
-!- parijat [75dc2d05@gateway/web/freenode/ip.117.220.45.5] has joined #shogun | 08:17 | |
parijat | iglesiasg: hi! | 08:17 |
parijat | iglesiasg: I just saw your comments. About using real_max | 08:18 |
parijat | iglesiasg: actually CHAID handles MISSING value as just another attribute value | 08:19 |
parijat | iglesiasg: so implementation-wise using MISSING as a valid float value gives me a lot of advantage | 08:20 |
@iglesiasg | parijat, ok | 08:21 |
parijat | iglesiasg: for example if you see merge method for ordinal attributes, there is nothing special done for missing attribute | 08:21 |
@iglesiasg | parijat, I trust your criteria :) | 08:21 |
parijat | iglesiasg: :) | 08:22 |
@iglesiasg | parijat, maybe I missed it, but did you mention in the doc that the user should use real_max to denote missing values? | 08:22 |
parijat | iglesiasg: nope not at the brief part! I should do it probably. I have mentioned it in the declaration place | 08:23 |
parijat | iglesiasg: This CHAID PR really got big without me realizing beforehand! | 08:24 |
-!- thor|home [~thoralf@91-65-141-142-dynip.superkabel.de] has joined #shogun | 08:26 | |
thor|home | Hello. | 08:26 |
@iglesiasg | all right | 08:26 |
@iglesiasg | gtg now, see you later guys! | 08:26 |
parijat | iglesiasg: bbye! | 08:27 |
parijat | :) | 08:27 |
@iglesiasg | parijat, let me know in the PR if you plan to update something or I should just merge | 08:27 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 08:27 | |
-!- parijat [75dc2d05@gateway/web/freenode/ip.117.220.45.5] has quit [Quit: Page closed] | 08:28 | |
-!- thor|home [~thoralf@91-65-141-142-dynip.superkabel.de] has quit [Quit: Konversation terminated!] | 08:59 | |
-!- lambday [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 240 seconds] | 08:59 | |
-!- lambday [67157f4c@gateway/web/freenode/ip.103.21.127.76] has joined #shogun | 09:02 | |
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun | 09:15 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 09:15 | |
lambday | HeikoS: hey! do you have a min? | 09:17 |
@HeikoS | lambday: hello! | 09:19 |
@HeikoS | lambday: how are things going? | 09:19 |
lambday | HeikoS: hey | 09:19 |
lambday | HeikoS: good actually | 09:19 |
@HeikoS | nice, good to hear! | 09:19 |
lambday | HeikoS: so I made a quick python implementation to check results against | 09:19 |
lambday | and things are sane | 09:19 |
lambday | HeikoS: with this new design, adding b-test would just take 15 mins | 09:20 |
@HeikoS | lambday: cool! very nice to hear that | 09:20 |
@HeikoS | lambday: yeah quick python implementation is they way to test against | 09:21 |
lambday | HeikoS: yeah mentioned the link in commit msg | 09:21 |
@HeikoS | lambday: and then put this into a notebook gist, send us for validation, and then put the link into unit test or so | 09:21 |
lambday | HeikoS: alright.. | 09:22 |
@HeikoS | lambday: looking very forward to play with this stuff btw! | 09:22 |
lambday | HeikoS: I'm just checking how many integration tests I broke btw :D | 09:22 |
lambday | HeikoS: yeah! here is the python code btw - https://github.com/lambday/shogun-hypothesis-testing/blob/master/btest.py | 09:22 |
lambday | HeikoS: you and Dino please check this once you get some free time... if this one is at least okay | 09:23 |
lambday | HeikoS: then its easier to verify shogun code | 09:23 |
lambday | made some changes later for generating unit-tests values :/ | 09:25 |
@HeikoS | lambday: yeah the unit tests are important. integration tests not so much for this one | 09:25 |
@HeikoS | lambday: but if you break them, make surre to look at the errors before re-generatiung | 09:26 |
@HeikoS | lambday: you know there is linspace in numpy? :) | 09:26 |
lambday | HeikoS: oh they have it already! | 09:26 |
lambday | argh I wrote creepy lambda stuff | 09:27 |
@HeikoS | lambday: code looks fine btw, gotta check the things in detail but good on first look | 09:27 |
lambday | HeikoS: okay.. let me know if its okay later when you get some free time | 09:28 |
lambday | HeikoS: meanwhile I'm fixing broken ingration tests and then will send PR | 09:28 |
@HeikoS | lambday: I am free now, what else is up? | 09:28 |
lambday | HeikoS: alright so I pushed the code to my fork already | 09:28 |
lambday | HeikoS: https://github.com/lambday/shogun/compare/feature;btest | 09:29 |
lambday | the unit-tests agrees with the results from the python code | 09:30 |
lambday | so, for other dependent stuffs, I used local machine values for unit-tests | 09:30 |
lambday | specially kernel selection unit-tests | 09:30 |
@HeikoS | lambday: okay | 09:34 |
@HeikoS | lambday: yeah I mean the unit tests so far were written this way | 09:34 |
@HeikoS | lambday: I just did stuff on my machine locally (from my MSc thesis code) and then put the numbers is, which is pretty bad | 09:35 |
@HeikoS | so it cant be worse ;) | 09:35 |
lambday | HeikoS: haha :D having a reference implementation is best.. I don't have any for kernel selection stuffs.. | 09:36 |
lambday | HeikoS: but I guess as long as basic statistic/var computation is verified with external code, other dependent stuffs follow from it - so it sort of works as before | 09:37 |
@HeikoS | lambday: yeah | 09:52 |
@HeikoS | lambday: so the kernel selection stuff worked fine | 09:52 |
@HeikoS | lambday: its very easy after all | 09:52 |
thoralf | Heyhey. | 09:53 |
@HeikoS | lambday: only the kernel combinations might be changed since they need all of the h-terms and their covariance | 09:53 |
@HeikoS | lambday: btw feel free to you your new awespome linalg backend for all spectral things :) | 09:53 |
@HeikoS | thoralf: hey! | 09:53 |
lambday | HeikoS: yeah - well, the covariance stuff is changed a bit in compute_statistic_and_Q .. basically what I did is to stream 2 blocks from both the distributions and computed online cov... no h-vectors are there anymore since statistic directly computes a scalar value blockwise in the new formula | 09:56 |
lambday | HeikoS: with the new implementation, the weights for in the kernel selection are *pretty* close to what it was before | 09:57 |
lambday | linalg - yeah! as soon as it makes it to develop :) | 10:01 |
-!- HeikoS [~heiko@90.201.233.229] has quit [Ping timeout: 255 seconds] | 10:04 | |
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun | 10:07 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:08 | |
@HeikoS | lambday: cool, that is good, they cannot be the same, but close is fine | 10:08 |
@HeikoS | lambday: more important that the statistic/variance is correct and the kernel selection procedure should do the rest, btw feel free to review this code a bit and point out weaknesses | 10:09 |
@HeikoS | lambday: did besser82 or wiking ever got back to you on the merge of linalg? | 10:09 |
lambday | HeikoS: umm.. bo | 10:09 |
lambday | no* | 10:09 |
lambday | HeikoS: btw khalednasr had some great ideas regarding using a wrapper class for viennacl matrices | 10:10 |
lambday | HeikoS: that was one big issue with the existing thing (remember why you asked to add at least a few more methods) :D | 10:11 |
lambday | but that's solvable using his ideas | 10:11 |
-!- HeikoS [~heiko@90.201.233.229] has quit [Quit: Leaving.] | 10:35 | |
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun | 11:14 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 11:14 | |
-!- lisitsyn [~qdrgsm@80.252.20.67] has quit [Ping timeout: 260 seconds] | 11:59 | |
@HeikoS | lambday: yeah I saw his suggestions. +1 from my side | 12:18 |
@HeikoS | lambday: I dont know where everyone is, but we should aim to merge this stuff soon. | 12:19 |
@HeikoS | lambday: what about some cholesky and triangular solvers? | 12:19 |
lambday | HeikoS: well, before that I think this viennacl matrix wrapper stuff should be finished | 12:20 |
lambday | so we'd get a clear idea whether we can use their matrices as well - since they too have many solvers | 12:20 |
lambday | if khaled's idea turns out to be great, then rest can be added more easily | 12:21 |
@HeikoS | lambday: I see, cool! | 12:21 |
lambday | HeikoS: man I've been experimenting with p-values with old and new... | 12:21 |
lambday | HeikoS: is it normal for linear time mmd to give higher type 1 errors? | 12:22 |
lambday | even in the previous implementation its pretty high | 12:22 |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has joined #shogun | 12:29 | |
@HeikoS | lambday: type1? | 12:45 |
@HeikoS | lambday: that depends on the quality of the asymptotic approximations to the null | 12:45 |
@HeikoS | how many samples? | 12:45 |
@HeikoS | you can always check that via sampling the null | 12:45 |
@HeikoS | and look at the histogram | 12:46 |
@HeikoS | if it looks very Gaussian, the type1 error should be exactly as desired | 12:46 |
lambday | HeikoS: I was working with existing examples only... tried 1000,10000,100000 samples with 0.5 mean shift | 12:46 |
lambday | yeah plot is better.. trying that | 12:47 |
@HeikoS | lambday: if the plot is good but the type1 is wrong, then there is a bug | 12:52 |
lambday | HeikoS: what ideally should be the correct type1 error? | 12:53 |
lambday | HeikoS: plot is pretty gaussian | 12:53 |
lambday | with mean around 0 | 12:53 |
@HeikoS | type1 error is asymptotically euqal to p-value | 13:11 |
lambday | HeikoS: I guess then it makes sense - I also tried changing the blocksize and as Dino mentioned in the writeup, this approach gives correct variance estimate when blocksize is small but overestimates when blocksize is large | 13:21 |
lambday | made some plots | 13:21 |
@HeikoS | cool! | 13:21 |
@HeikoS | lambday: put them into a notebook, we can re-use this later | 13:21 |
lambday | HeikoS: alright | 13:22 |
lambday | HeikoS: data is going to change for integration tests - I'm first checking whether the changes make sense... then sending pr | 13:23 |
@HeikoS | lambday: yeah, this is fine, just monitor whether things are like totally different | 13:24 |
@HeikoS | but on the other hand if everything is probably unit tested then we should be fine | 13:24 |
lambday | HeikoS: yeah its all unit-tested... | 13:24 |
-!- lambday [67157f4c@gateway/web/freenode/ip.103.21.127.76] has quit [Ping timeout: 240 seconds] | 14:48 | |
-!- HeikoS [~heiko@90.201.233.229] has quit [Ping timeout: 252 seconds] | 14:52 | |
-!- Saurabh7 [~Saurabh7@117.203.115.189] has joined #shogun | 15:09 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 15:56 | |
-!- HeikoS [~heiko@90.201.233.229] has joined #shogun | 16:09 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:09 | |
-!- thor|home [~thoralf@46.114.135.36] has joined #shogun | 16:29 | |
thor|home | Heyho. | 16:30 |
-!- kislay [~Abhijeet@103.240.205.41] has joined #shogun | 16:44 | |
-!- thor|home [~thoralf@46.114.135.36] has quit [Ping timeout: 240 seconds] | 16:57 | |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 17:43 | |
shogun-notifier- | shogun: Parijat Mazumdar :develop * 1bc0876 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/1bc0876affb66b686c741d02dcf499a7d9815eaa | 17:43 |
shogun-notifier- | shogun: minor changes in CART | 17:43 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 89ccf88 / / (3 files): https://github.com/shogun-toolbox/shogun/commit/89ccf88347c060602bf262543074265de9e653f8 | 17:43 |
shogun-notifier- | shogun: Merge pull request #2276 from mazumdarparijat/CART | 17:43 |
shogun-notifier- | shogun: | 17:43 |
shogun-notifier- | shogun: minor changes in CART | 17:43 |
-!- travis-ci [~travis-ci@ec2-54-197-181-124.compute-1.amazonaws.com] has joined #shogun | 17:49 | |
travis-ci | [travis-ci] it's khalednasr's turn to pay the next round of drinks for the massacre he caused in khalednasr/shogun: http://travis-ci.org/khalednasr/shogun/builds/26671140 | 17:49 |
-!- travis-ci [~travis-ci@ec2-54-197-181-124.compute-1.amazonaws.com] has left #shogun [] | 17:49 | |
shogun-buildbot_ | build #658 of precise - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/precise%20-%20libshogun/builds/658 | 18:10 |
-!- travis-ci [~travis-ci@ec2-54-197-181-124.compute-1.amazonaws.com] has joined #shogun | 18:23 | |
travis-ci | [travis-ci] it's Fernando Iglesias's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/26677826 | 18:23 |
-!- travis-ci [~travis-ci@ec2-54-197-181-124.compute-1.amazonaws.com] has left #shogun [] | 18:23 | |
-!- pickle27f [~pickle27@216.191.231.218] has quit [Quit: This computer has gone to sleep] | 18:27 | |
-!- pickle27f [~pickle27@216.191.231.218] has joined #shogun | 18:36 | |
-!- Saurabh7 [~Saurabh7@117.203.115.189] has quit [Quit: Leaving] | 18:54 | |
shogun-buildbot_ | build #2346 of deb3 - modular_interfaces is complete: Failure [failed test libshogun] Build details are at http://buildbot.shogun-toolbox.org/builders/deb3%20-%20modular_interfaces/builds/2346 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, Parijat Mazumdar <mazumdarparijat@gmail.com> | 18:56 |
-!- soumyaC [uid15286@gateway/web/irccloud.com/x-ifzwtgtcgpzqolkq] has joined #shogun | 19:18 | |
-!- khalednasr [~k.nasr92@41.69.147.3] has joined #shogun | 19:19 | |
-!- HeikoS [~heiko@90.201.233.229] has quit [Ping timeout: 252 seconds] | 19:24 | |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 19:41 | |
-!- HeikoS [~heiko@nat-164-250.internal.eduroam.ucl.ac.uk] has joined #shogun | 19:58 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:58 | |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 20:43 | |
-!- zxtx [~zv@98.223.196.32] has quit [Ping timeout: 240 seconds] | 20:46 | |
-!- thor|home [~thoralf@2a02:2028:1d7:1711:840b:45d4:2899:f55] has joined #shogun | 20:48 | |
-!- kislay [~Abhijeet@103.240.205.41] has quit [Quit: Leaving] | 20:59 | |
-!- thor|home [~thoralf@2a02:2028:1d7:1711:840b:45d4:2899:f55] has quit [Ping timeout: 265 seconds] | 21:37 | |
-!- HeikoS [~heiko@nat-164-250.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 21:40 | |
-!- thor|home [~thoralf@2a02:2028:1d7:1711:840b:45d4:2899:f55] has joined #shogun | 21:41 | |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has joined #shogun | 21:47 | |
-!- thor|home [~thoralf@2a02:2028:1d7:1711:840b:45d4:2899:f55] has left #shogun ["Konversation terminated!"] | 21:51 | |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has joined #shogun | 22:13 | |
-!- zxtx [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Remote host closed the connection] | 22:13 | |
-!- zxtx_ [~zv@129-79-241-150.dhcp-bl.indiana.edu] has joined #shogun | 22:14 | |
-!- zxtx_ [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Quit: Leaving] | 22:42 | |
-!- zxtx_ [~zv@129-79-241-150.dhcp-bl.indiana.edu] has joined #shogun | 22:43 | |
-!- khalednasr [~k.nasr92@41.69.147.3] has quit [Quit: Leaving] | 22:56 | |
--- Log closed Tue Jun 03 23:13:53 2014 | ||
--- Log opened Tue Jun 03 23:14:01 2014 | ||
-!- shogun-toolbox [~shogun@7nn.de] has joined #shogun | 23:14 | |
-!- Irssi: #shogun: Total of 16 nicks [3 ops, 0 halfops, 0 voices, 13 normal] | 23:14 | |
-!- Irssi: Join to #shogun was synced in 7 secs | 23:14 | |
-!- zxtx_ [~zv@129-79-241-150.dhcp-bl.indiana.edu] has quit [Ping timeout: 265 seconds] | 23:14 | |
-!- PirosB3 [~pirosb3@host80-223-dynamic.246-95-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 23:23 | |
--- Log closed Wed Jun 04 00:00:17 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!