--- Log opened Mon Oct 29 00:00:00 2012 | ||
--- Day changed Mon Oct 29 2012 | ||
blackburn | yes | 00:00 |
---|---|---|
wiking | :P | 00:00 |
heiko | wiking, btw I just read the word unit test, whats the state of that? | 00:00 |
blackburn | it have longed a month already | 00:00 |
blackburn | has longed :D | 00:00 |
blackburn | I can haz | 00:00 |
@sonney2k | blackburn, uhh ranking algorithms | 00:00 |
@sonney2k | I've talked to the guy who did that at yahoo | 00:00 |
@sonney2k | research | 00:00 |
blackburn | sonney2k: I don't care what to do for that money | 00:01 |
blackburn | :D | 00:01 |
wiking | heiko: lol good grep. well the thing is that it's tricky. i have to add extra code into the makefile so that gmock and gtest are compiled with the same flags as shogun. and u need to provide the src of gmock/gtest to shogun to be able to do that | 00:01 |
@sonney2k | blackburn, learning to rank is certainly interesting | 00:01 |
wiking | heiko: lot of mindless ./configure and Makefile hacking | 00:01 |
blackburn | if being serious it is researchy | 00:01 |
blackburn | my current job is like a factory worker job | 00:01 |
blackburn | and twice less money | 00:01 |
@sonney2k | twice less == half ? | 00:02 |
wiking | blackburn: clearcase!!!! | 00:02 |
blackburn | haha yes | 00:02 |
heiko | wiking, so one downloads gmock/gtest source, and provides path to shogun configure and then everything is done automatically? | 00:02 |
wiking | blackburn: ain't no factory without clearcase :D | 00:02 |
@sonney2k | wiking, what is clearcase? | 00:02 |
blackburn | wiking: ain't no sunshine when clearcase | 00:02 |
heiko | wiking, maybe one could even add a script that downloads the google libs via wget or so? | 00:03 |
blackburn | know that song? | 00:03 |
blackburn | :D | 00:03 |
blackburn | sonney2k: cvs | 00:03 |
wiking | heiko: then u compile those sources with the same flags as shogun and then u have your static libs that u link it to our unittesting objects | 00:03 |
wiking | heiko: and then u are done :) | 00:03 |
wiking | heiko: well actually standard distribs provide them | 00:03 |
heiko | ok, wiking, does ubuntu? | 00:04 |
heiko | there were these problems I remember | 00:04 |
wiking | heiko: gtest yes | 00:04 |
wiking | heiko: gmock no | 00:04 |
wiking | :) | 00:04 |
heiko | k | 00:04 |
wiking | gmock is still a precompiled library :P | 00:04 |
blackburn | sonney2k: when one commits in git other guy in clearcase creates activity, check outs, ...... , commits, check ins, closes activity | 00:04 |
wiking | but in case of gtest it puts the source under /usr/src.. | 00:04 |
heiko | Maybe when you are done, you could write a small howto and put it in the test directory | 00:04 |
wiking | sonney2k: that is clearcase ^ :D | 00:04 |
blackburn | and during asks for proper rights, tries to grasp its interface and attempts to kill himself | 00:05 |
wiking | heiko: aintnotime :D | 00:05 |
wiking | heiko: will do after next week maybe | 00:05 |
heiko | wiking, no worries :) I know the problem | 00:05 |
heiko | wiking, what are you up to currenlty?P | 00:05 |
wiking | heiko: paper | 00:05 |
wiking | :> | 00:05 |
wiking | and trying to get within the deadline :D | 00:05 |
@sonney2k | heiko, the new doc should appear if tonights nightly_default builds | 00:06 |
heiko | sonney2k, cool thing! whats the link? | 00:06 |
@sonney2k | http://shogun-toolbox.org/doc/en/current/shogun_tutorial.pdf | 00:06 |
@sonney2k | actually latest | 00:06 |
@sonney2k | not current | 00:06 |
@sonney2k | current doesn't have that | 00:06 |
@sonney2k | heiko, feel free to update the website! | 00:07 |
@sonney2k | with pointers... | 00:07 |
blackburn | can't open sonney2k | 00:07 |
heiko | sonney2k, ok Ill have a try | 00:07 |
wiking | mm china will dist-upgrade next month :) | 00:07 |
@sonney2k | blackburn, heiko well *after* the build tonight | 00:07 |
blackburn | ahh | 00:08 |
heiko | sonney2k, sure | 00:08 |
-!- shogun-buildbot [~shogun-bu@7nn.de] has quit [Quit: buildmaster reconfigured: bot disconnecting] | 00:08 | |
-!- shogun-buildbot [~shogun-bu@7nn.de] has joined #shogun | 00:08 | |
heiko | btw I just wrote a bit more on the MMD stuff | 00:08 |
heiko | we should keep on doing this - very nice thing to have | 00:08 |
heiko | blackburn, I also added a part that explains SHOGUNs structures, like Features, Kernels, etc | 00:08 |
blackburn | heiko: nice | 00:09 |
heiko | And I want to add a part about the model selection soon | 00:09 |
@sonney2k | heiko, I can basically copy all kernels form my phd thesis | 00:09 |
@sonney2k | so kernel section should be easy | 00:09 |
@sonney2k | coffin from some other paper | 00:09 |
heiko | and part of my job will be documentation, so the GP part will be documented soon also | 00:09 |
@sonney2k | lets hope it gets accepted | 00:09 |
heiko | sonney2k, cool! add that | 00:09 |
@sonney2k | but heiko, last year only orgs of the size of KDE got in | 00:09 |
heiko | sonney2k, really? ok - but I mean we are growing and have quite a community | 00:10 |
heiko | also we need documentation more than anybody else | 00:10 |
heiko | if a Desktop system is not self-explaining it has failed anyways | 00:10 |
wiking | heiko: the 5 or 6 of us u mean? ;) | 00:10 |
heiko | wiking, :D | 00:10 |
@sonney2k | shogun-buildbot, force build nightly_default | 00:10 |
shogun-buildbot | The build has been queued, I'll give a shout when it starts | 00:11 |
blackburn | heiko: growing? | 00:11 |
blackburn | who's new? :D | 00:11 |
heiko | wiking, sonney2k, blackburn, actually, I do quite some PR for shogun around here, but people really often complain about documentation and that thats the reason they dont use it | 00:11 |
@sonney2k | I can believe so | 00:11 |
heiko | blackburn, the 2012 GSOC student for example | 00:11 |
blackburn | heiko: wiking? | 00:12 |
heiko | blackburn, yes for example :) | 00:12 |
@sonney2k | if n4nd0 wasn't sleeping so much he would be here too I guess :D | 00:12 |
@sonney2k | which reminds me | 00:12 |
@sonney2k | I wanted to ask him to put his example in applications | 00:12 |
wiking | ? | 00:12 |
blackburn | wiking: pass ;) | 00:12 |
wiking | :D | 00:12 |
blackburn | NOP | 00:13 |
wiking | nopnop | 00:13 |
wiking | :> | 00:13 |
heiko | speaking of sleeping, I will go to bed now, see you all! | 00:13 |
wiking | nite | 00:13 |
-!- heiko [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has left #shogun [] | 00:14 | |
blackburn | sonney2k: wiking did I told you I am now mentoring at my current job? | 00:15 |
@sonney2k | indeed seems like fish got used to the tank - so I rather sleep now too | 00:15 |
@sonney2k | blackburn, heh yes | 00:15 |
blackburn | sonney2k: apparently one my student disappeared | 00:15 |
blackburn | :D | 00:15 |
@sonney2k | blackburn, maybe you should not mentor shogun then ;-) | 00:16 |
wiking | blackburn: noup, whom/what? :) | 00:16 |
@sonney2k | with these fish it is tough to check if they are all around - they are pretty 2d | 00:16 |
@sonney2k | so from the wrong angle hard to recognize | 00:16 |
blackburn | wiking: they gave me 2 students to mentor | 00:16 |
blackburn | funny that one (not disappeared one) is of same age as me | 00:16 |
blackburn | but knows no java no patterns nothing | 00:17 |
wiking | blackburn: that happens... | 00:17 |
wiking | "Little is known about his personal life, beyond a liking for basketball and, according to a leaked US diplomatic cable, Hollywood war movies, possibly acquired during a brief stay in Iowa when he was a young man." | 00:17 |
wiking | hehehe | 00:17 |
wiking | i bet this guy has all the chuck norris movies :D | 00:17 |
blackburn | wiking: who it is about? | 00:17 |
wiking | blackburn: Xi Jinping | 00:17 |
wiking | the distupgrade guy for china | 00:18 |
wiking | :) | 00:18 |
wiking | blackburn: don't u read other stuff than slashdot? :D | 00:18 |
blackburn | ahh I am not a big fan of china | 00:18 |
blackburn | north korea is my love | 00:18 |
blackburn | wiking: I don't | 00:18 |
wiking | :> | 00:18 |
wiking | blackburn: pravda? :) | 00:19 |
blackburn | wiking: da | 00:19 |
blackburn | wiking: how much time would you able to speak in russian with me? :D | 00:19 |
wiking | nikoliko? :) | 00:20 |
wiking | mislim malo bi mi razumeo ako bi hteo | 00:20 |
wiking | :D | 00:20 |
blackburn | wiking: no idea :D | 00:21 |
wiking | ali russi nekako ne razumeju sprski tako dobro ka srbi razumeju ruski | 00:21 |
wiking | e pa tako | 00:21 |
wiking | ;D | 00:21 |
blackburn | ahh | 00:21 |
wiking | NU PA GADJI!!! | 00:21 |
blackburn | wiking: I understand kinda 70% | 00:21 |
blackburn | :D | 00:21 |
wiking | hehhehe | 00:21 |
wiking | http://www.youtube.com/watch?v=5B9XYZqYuN8 | 00:22 |
wiking | some oldschool stuff :> | 00:22 |
blackburn | how russians do not understand srpska at all | 00:22 |
wiking | u know nu pagadi? | 00:22 |
wiking | blackburn: yep | 00:22 |
blackburn | so do serbian | 00:22 |
wiking | vidis razumes malo ;) | 00:22 |
blackburn | wiking: come on how can I do not know nu pogodi | 00:22 |
wiking | hehe | 00:24 |
wiking | blackburn: well u've born after soviet union | 00:25 |
blackburn | wiking: before! | 00:25 |
blackburn | :D | 00:25 |
wiking | ?! | 00:25 |
wiking | noway | 00:25 |
blackburn | march 1991 | 00:25 |
wiking | gorbachov already fixed thing by then | 00:25 |
blackburn | yes | 00:25 |
wiking | so u were born in russia | 00:25 |
wiking | not in soviet union | 00:25 |
wiking | :) | 00:25 |
blackburn | I was born in soviet union | 00:26 |
blackburn | do not mix things :D | 00:26 |
blackburn | soviet union disappeared in december | 00:26 |
blackburn | I was born in march | 00:26 |
wiking | hahaha | 00:26 |
wiking | couple of months | 00:26 |
wiking | doesn't count :P | 00:26 |
blackburn | counts! | 00:26 |
blackburn | :D | 00:26 |
wiking | so yeah this is like true soviet propaganda cartoon | 00:27 |
wiking | love this shit :D | 00:27 |
blackburn | hey no propaganda at all | 00:27 |
blackburn | it is very kind | 00:27 |
wiking | no fucking way: The latest episode was produced in 2006 | 00:28 |
wiking | really?! | 00:28 |
wiking | i thought that it was only produced like in 1970's | 00:28 |
blackburn | wiking: well these latest things are not that cool | 00:30 |
blackburn | wiking: I'd say it is soviet propaganda as much as tom&jerry is american propaganda :) | 00:30 |
shogun-buildbot | build #149 forced | 00:31 |
shogun-buildbot | I'll give a shout when the build finishes | 00:31 |
wiking | blackburn: well t&j is like full on american dream propaganda | 00:31 |
wiking | :D | 00:31 |
blackburn | wiking: okay then I'd say it contains much less propaganda | 00:32 |
blackburn | :) | 00:32 |
blackburn | I can't really see anything really soviet | 00:32 |
@sonney2k | nite | 00:42 |
blackburn | sonney2k: good night | 00:43 |
blackburn | wiking: have you ever heard john zorn's music? | 00:50 |
shogun-buildbot | build #149 of nightly_default is complete: Failure [failed tutorial] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/149 | 01:13 |
-!- blackburn [~blackburn@188.168.3.27] has quit [Quit: Leaving.] | 01:36 | |
shogun-buildbot | build #150 of nightly_default is complete: Failure [failed tutorial] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/150 | 04:00 |
-!- ptizoom [~christian@79-71-87-88.dynamic.dsl.as9105.com] has quit [Quit: Ex-Chat] | 04:32 | |
-!- ptizoom_ [~christian@79-71-87-88.dynamic.dsl.as9105.com] has joined #shogun | 04:32 | |
-!- ptizoom_ [~christian@79-71-87-88.dynamic.dsl.as9105.com] has quit [Ping timeout: 268 seconds] | 05:02 | |
-!- ptizoom_ [~christian@85.210.80.20] has joined #shogun | 05:15 | |
-!- blackburn [~blackburn@83.234.54.215] has joined #shogun | 06:57 | |
-!- blackburn [~blackburn@83.234.54.215] has quit [Quit: Leaving.] | 07:27 | |
-!- ptizoom_ [~christian@85.210.80.20] has quit [Read error: Connection reset by peer] | 08:22 | |
-!- ptizoom__ [~christian@85.210.80.20] has joined #shogun | 08:22 | |
-!- ptizoom__ is now known as ptizoom | 08:53 | |
-!- sonne|work [~sonnenbu@194.78.35.195] has joined #shogun | 09:19 | |
-!- heiko [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has joined #shogun | 09:31 | |
sonne|work | heiko: seen that the pdflatex command failed? | 09:33 |
sonne|work | what extra fonts do we need? | 09:33 |
heiko | sonne|work, havent seen, let me check | 09:34 |
heiko | sonne|work another question, do you know when approx. we will get notified whether we are invited to the doc summit? | 09:34 |
sonne|work | in about a week | 09:34 |
heiko | sonne|work, is texlive-fonts-recommended installed? | 09:39 |
sonne|work | no idea | 09:40 |
sonne|work | cannot check from here | 09:40 |
heiko | sonne|work, I think thats it, at least in ubuntu | 09:41 |
heiko | sonne|work, about the doc summit, any guess about our chances? The thing is I would have to fly to Germany and apply for an Express Passport and thats expensive so I would only like to do it if needed. My NIPS travel funding application got neglected so if doc summit doesnt work then I dont need to get the quick passport | 09:44 |
sonne|work | heiko: well I cannot say more - carol said they want to be quick | 10:05 |
sonne|work | so hopefully this week | 10:05 |
-!- heiko [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has quit [Disconnected by services] | 10:34 | |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has joined #shogun | 10:35 | |
-!- yoo__ [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has joined #shogun | 10:37 | |
-!- yoo__ [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has quit [Ping timeout: 245 seconds] | 10:45 | |
-!- yoo [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has joined #shogun | 10:48 | |
yoo | hi all | 10:48 |
sonne|work | heiko1: I wonder why you didn't get selected for nips travel funding. having a paper there used to be enough... | 11:02 |
sonne|work | hi yoo | 11:03 |
heiko1 | sonne|work - no idea :( | 11:14 |
sonne|work | I remember having received this grant a couple of times | 11:15 |
sonne|work | but might be that I had an oral (spotlight!) | 11:15 |
sonne|work | or some mixtures of papers, demo & workshop talks | 11:15 |
heiko1 | I submitted to a workshop, but these havent been decided yet | 11:19 |
heiko1 | sonne|work, I mean its fine, there will be other times | 11:19 |
heiko1 | also, once I am in a phd programe, it will be much easier | 11:20 |
heiko1 | especially here, people are quite well funded | 11:20 |
heiko1 | yoo, is the example that you used the same I sent in the pastebin? | 11:23 |
heiko1 | or did you change something, | 11:23 |
heiko1 | because it worked here | 11:24 |
yoo | no changes, but you have to change kernels .. | 11:24 |
yoo | Commwordstringker works fine | 11:24 |
yoo | the prob I reported is with the weightedcommstringker | 11:24 |
heiko1 | ok | 11:25 |
heiko1 | parameters? | 11:25 |
yoo | in my example one of the two kernels was commented | 11:25 |
yoo | same | 11:25 |
yoo | http://pastebin.com/zsKQYPhM | 11:26 |
sonne|work | heiko1: well you have a nips paper w/o being a phd student. that is already quite something! | 11:26 |
heiko1 | sonne|work, I know - I would have loved to go there, well lets see | 11:28 |
heiko1 | the thing is: the doc summit is in parallel | 11:28 |
heiko1 | so there might be still a chance :) but I am not too optimistic about that | 11:29 |
heiko1 | yoo, confirmed, I get the same error | 11:29 |
heiko1 | yoo, I will have a quick look now | 11:29 |
yoo | one of the features get order = 0 and this cause the error | 11:30 |
heiko1 | yoo, looks like a X-Val based error to mee | 11:31 |
heiko1 | with subsets on the features not updating the order | 11:32 |
heiko1 | sonne|work, the get_order() of string features, what does it do? does it change when a subset is applied? | 11:32 |
sonne|work | no it should stay fixed | 11:32 |
sonne|work | heiko1: basically order == n (from n-gram length) | 11:33 |
sonne|work | so substrings of length n are considered (but this is coded in the features already) | 11:34 |
heiko1 | how can this get zero then? | 11:34 |
heiko1 | in x-val | 11:34 |
yoo | yes it occurs only with xval | 11:35 |
yoo | weightedcommwordstringker manage several order right ? | 11:36 |
heiko1 | yoo, I actually have no idea about this kernel, never looked at it, sonne2y aka sonne|work is the specialist here | 11:36 |
heiko1 | yoo, but I think the bug comes from somewhere else | 11:36 |
heiko1 | let me investigate a bit further | 11:36 |
sonne|work | yoo: yes spectrum kernel using all orders up to order | 11:37 |
heiko1 | sonne|work, wow didnt know that I thought this was called the blended spectrum kernel, good to know that this is there :) | 11:39 |
sonne|work | everyone re-invents the wheel so we have multiple names | 11:39 |
sonne|work | heiko1: but yes that is the one. caution though it needs special args in the obtain_from_string / translate function (see example) | 11:40 |
heiko1 | sonne|work, I see, this is exactly where erics bug comes from: store_model_features calls copy_subset which copies a certain part of the data and creates a new stringfeatures instance, but does not call this obtain from char | 11:41 |
heiko1 | yoo, so problem solbed, but another one pops up ... :) | 11:45 |
sonne|work | heiko1: hmmhh I don't know - this translation is only done once | 11:46 |
sonne|work | so all subsets should be fine | 11:46 |
heiko1 | sonne|work, but this is different | 11:46 |
heiko1 | its store_model_features | 11:46 |
heiko1 | so it creates a copy of the features object where only a certain subset of the data is included | 11:46 |
sonne|work | heiko1: yes but still | 11:46 |
heiko1 | for example, order is not even set | 11:47 |
heiko1 | in the new instance | 11:47 |
sonne|work | it should then just copy order and everything and all good | 11:47 |
heiko1 | yes did, but now symbol_mask_table is missing | 11:47 |
heiko1 | so I was wondering whether there was a more systematic way of just copiying till no more assertion errors | 11:47 |
heiko1 | since I might miss something this way | 11:47 |
heiko1 | which chages results or so... | 11:47 |
heiko1 | what about an embed_features call? | 11:50 |
heiko1 | that seems to do it | 11:50 |
heiko1 | sonne|work, hope that is fine? | 11:50 |
sonne|work | what would it do? | 11:51 |
heiko1 | I now even did differently | 11:52 |
heiko1 | result->order=order; | 11:52 |
heiko1 | result->compute_symbol_mask_table(result->alphabet->get_num_symbols()); | 11:52 |
heiko1 | then, at least the code runs through | 11:52 |
heiko1 | but I dont know if I changed any semantics this way | 11:53 |
heiko1 | yoo, its fixed, submitting patch soon | 11:56 |
heiko1 | yoo, maybe you can have a look if the results make sense, I am not too sure whether this fix is correct in terms of that data is not changed or something | 11:57 |
yoo | heiko1: thx allright, I will test it. I am trying to figure out what was the pb, could you explain ? | 11:57 |
heiko1 | yoo, the problem was realted to cross.validation | 11:58 |
heiko1 | kernel machines have to store the data for the model somehow | 11:58 |
heiko1 | but since we are working on subsets, these got changed again when testing on different sets in x-val | 11:58 |
heiko1 | thats why they store copies of the data once they computed the alpha terms | 11:59 |
heiko1 | and this method wasnt keeping the order for string features | 11:59 |
heiko1 | Ill send you the patch, its just two lines | 11:59 |
yoo | okay | 11:59 |
heiko1 | yoo, pretty hard to find if you dont know the framework and how cross-validation in shogun works internally | 11:59 |
yoo | good exercice then .. | 12:00 |
-shogungit:#shogun- [shogun] karlnapf opened pull request #809: fixes a bug (master...master) https://github.com/shogun-toolbox/shogun/pull/809 | 12:01 | |
heiko1 | yoo, https://github.com/shogun-toolbox/shogun/pull/809 | 12:01 |
-shogungit:#shogun- [shogun] karlnapf pushed 3 new commits to master: https://github.com/shogun-toolbox/shogun/compare/915af5cda5ab...1e93f0dd1c1b | 12:01 | |
-shogungit:#shogun- shogun/master 30d6b8a Heiko Strathmann: bugfix that allows using string features in cross-validation from store_model_features(). This method did not copy all attributed properly. | 12:01 | |
-shogungit:#shogun- shogun/master debc3a4 Heiko Strathmann: replaced print by debug message | 12:01 | |
-shogungit:#shogun- shogun/master 1e93f0d Heiko Strathmann: Merge pull request #809 from karlnapf/master... | 12:01 | |
heiko1 | now the code runs with the other kernel, Ill also write to the mailing list in case somebody else had this | 12:01 |
heiko1 | yoo, whats your name to give bug-hunting credits in the Changelog? :) | 12:02 |
heiko1 | I just write Eric (yoo) | 12:03 |
-shogungit:#shogun- [shogun] karlnapf opened pull request #810: mention bugfix in NEWS (master...master) https://github.com/shogun-toolbox/shogun/pull/810 | 12:04 | |
-shogungit:#shogun- [shogun] karlnapf pushed 2 new commits to master: https://github.com/shogun-toolbox/shogun/compare/1e93f0dd1c1b...91ea878b3d20 | 12:04 | |
-shogungit:#shogun- shogun/master 801071c Heiko Strathmann: mentioned last bugfix | 12:04 | |
-shogungit:#shogun- shogun/master 91ea878 Heiko Strathmann: Merge pull request #810 from karlnapf/master... | 12:04 | |
heiko1 | gotta go now, see you later! | 12:04 |
yoo | ++ | 12:09 |
-!- yoo [2eda6d52@gateway/web/freenode/ip.46.218.109.82] has quit [Quit: Page closed] | 12:09 | |
-shogungit:#shogun- [shogun] vigsterkr opened pull request #811: Minor fixes in LatentSVM (master...master) https://github.com/shogun-toolbox/shogun/pull/811 | 12:13 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has joined #shogun | 12:23 | |
sonne|work | wiking: you sure with that patch? | 13:08 |
-shogungit:#shogun- [shogun] sonney2k pushed 2 new commits to master: https://github.com/shogun-toolbox/shogun/compare/91ea878b3d20...528ccfd8a5cf | 13:09 | |
-shogungit:#shogun- shogun/master 77332c1 Viktor Gal: Minor fixes in LatentSVM | 13:09 | |
-shogungit:#shogun- shogun/master 528ccfd Soeren Sonnenburg: Merge pull request #811 from vigsterkr/master... | 13:09 | |
sonne|work | ok sure :D | 13:09 |
wiking | yes | 13:17 |
wiking | why? | 13:17 |
sonne|work | wiking: never mind | 13:33 |
sonne|work | X^{-1}^{-1}^{-1} problem | 13:34 |
blackburn | sonne|work: that must be pretty unstable to compute such matrix :) | 13:35 |
sonne|work | especially when neuron spikes are creating the data | 13:35 |
blackburn | sonne|work: is it possible to hold a code sprint at some conference? | 13:36 |
sonne|work | not officially and usually one is *very* busy at a conference | 13:36 |
blackburn | sonne|work: I believe we need some experts advice or so | 13:37 |
sonne|work | then we better get the experts to work with us | 13:38 |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has quit [Quit: Leaving.] | 13:40 | |
blackburn | sonne|work: yes, how? | 13:43 |
sonne|work | ask them :D | 13:43 |
blackburn | sonne|work: so I write a letter to vapnik okay | 13:43 |
blackburn | :D | 13:43 |
blackburn | so he ask his students of students of students to use shogun | 13:44 |
sonne|work | blackburn: he is retired :D | 13:44 |
blackburn | sonne|work: no he keeps pushing some his stuff | 13:44 |
blackburn | with lectures etc | 13:44 |
blackburn | sonne|work: just recently he was at MIT | 13:44 |
blackburn | chiyuan posted a few photos | 13:45 |
sonne|work | blackburn: where?\ | 13:45 |
blackburn | sonne|work: where what? | 13:45 |
sonne|work | photos | 13:45 |
blackburn | sonne|work: http://instagram.com/p/QIFbhhzHQ1/ | 13:46 |
blackburn | sonne|work: http://instagram.com/p/QDZKOBzHZd/ | 13:46 |
blackburn | vlad is more than full of energy :D | 13:46 |
blackburn | just 76 years | 13:47 |
blackburn | sonne|work: my supervisor was born just nn the same day our little war got over, very active still | 13:49 |
sonne|work | well yeah I've seen his talk a few times - he really is into his stuff still | 13:49 |
blackburn | 9 may 1945 :) | 13:49 |
blackburn | sonne|work: you have seen that for real? | 13:49 |
blackburn | I think I'd really like to listen to him | 13:50 |
wiking | blackburn: http://videa.hu/videok/jarmuvek/szetrobbant-a-lada-megsem-halt-qu5Z2ey02o7lIBa2 | 13:56 |
-!- gsomix [~gsomix@188.168.13.104] has joined #shogun | 13:57 | |
gsomix | hi | 13:57 |
blackburn | wiking: 2105 FTW | 14:01 |
wiking | :) | 14:02 |
sonne|work | blackburn: yes I have seen him talking about god einstein and himself :D | 14:16 |
sonne|work | blackburn: and he was very upset when oliver said sth. like you don't need kernels :) | 14:17 |
blackburn | sonne|work: don't need kernels for what? | 14:17 |
sonne|work | SVMs | 14:17 |
sonne|work | and no need to do anything in the dual | 14:18 |
sonne|work | haha | 14:18 |
blackburn | he meant linear primal svm? | 14:18 |
sonne|work | but olivier will mentor next year so we won ;) | 14:18 |
blackburn | sonne|work: what he wants to mentor? | 14:19 |
sonne|work | yeah. this guy http://www.kyb.mpg.de/fileadmin/user_upload/files/publications/attachments/primal_%5b0%5d.pdf | 14:19 |
sonne|work | blackburn: well I talked to him - I told you that helps at times ;) | 14:19 |
blackburn | ah so chapelle | 14:20 |
blackburn | he is an author of some SVM IIRC? | 14:20 |
blackburn | sonne|work: sometimes I feel SVM is stupid | 14:21 |
blackburn | :D | 14:21 |
sonne|work | replace SVM with $METHOD | 14:21 |
blackburn | sonne|work: we are all too tied to SVM | 14:21 |
blackburn | sonne|work: it seems to be so bad we have no methods but SVM | 14:22 |
wiking | yeah so heiko will make us GP classifiers :) | 14:22 |
blackburn | wiking: they are not better I heard | 14:23 |
sonne|work | blackburn: what is ``better'' then? | 14:23 |
wiking | blackburn: complexity is high but it's autotuning | 14:23 |
blackburn | sonne|work: no idea - accuracy-wise may be | 14:23 |
wiking | so you dont really need to tune your parameters | 14:23 |
wiking | but fuckthisshit | 14:24 |
blackburn | wiking: autotuning is interesting but performance? | 14:24 |
wiking | why my code is not working argh | 14:24 |
blackburn | can one do large scale? | 14:24 |
wiking | as said earlier | 14:24 |
wiking | high complexity | 14:24 |
blackburn | complexity? algorithmic or like in VC? | 14:24 |
wiking | o(n^3) | 14:24 |
wiking | if i remember right | 14:24 |
blackburn | wiking: fuckthishit | 14:24 |
wiking | hehehe | 14:25 |
wiking | well getsome/loose some | 14:25 |
wiking | as usual | 14:25 |
blackburn | and we came that everything is useless again | 14:25 |
blackburn | :D | 14:25 |
wiking | well if u give a neural net enough data and time, it'll learn god of course :) | 14:25 |
sonne|work | wiking, blackburn - matrix inversion! | 14:26 |
wiking | where? :D | 14:26 |
sonne|work | *EXPENSIVE* | 14:26 |
wiking | is it coming :) | 14:26 |
sonne|work | GPs | 14:26 |
blackburn | sonne|work: n^3/3 at least yeah | 14:26 |
* wiking hides from matrix inversions :D | 14:26 | |
sonne|work | and as blackburn said - rock stable stuff | 14:26 |
sonne|work | even more stable than QP optimizers | 14:27 |
sonne|work | haha | 14:27 |
blackburn | sonne|work: I am solving problems with det=1e-9 here come on | 14:27 |
blackburn | :D | 14:27 |
wiking | :>>> | 14:27 |
blackburn | if we are talking about illposed problems dimension reduction that is | 14:27 |
sonne|work | blackburn: at least positive :) | 14:27 |
blackburn | sonne|work: not always, IEEE754 FTW | 14:28 |
blackburn | heh | 14:28 |
blackburn | I add 1e-9 to diag | 14:28 |
blackburn | to avoid negative | 14:28 |
sonne|work | even the gaussian kernel can be negative w/ floating point numbers | 14:28 |
wiking | i've heard that in new york they are planning to build a sandy-bridge | 14:41 |
wiking | hehehehe | 14:41 |
wiking | :) | 14:41 |
wiking | i know it was a bad one... :P | 14:42 |
blackburn | heh | 14:45 |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has joined #shogun | 14:45 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 14:57 | |
-!- gsomix [~gsomix@188.168.13.104] has quit [Ping timeout: 255 seconds] | 15:23 | |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has quit [Quit: Leaving.] | 15:28 | |
-!- gsomix [~gsomix@188.168.13.104] has joined #shogun | 15:42 | |
-!- gsomix [~gsomix@188.168.13.104] has quit [Ping timeout: 260 seconds] | 15:49 | |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has joined #shogun | 15:59 | |
-!- gsomix [~gsomix@188.168.13.104] has joined #shogun | 16:21 | |
-!- blackburn [5bdfb203@gateway/web/freenode/ip.91.223.178.3] has quit [Quit: Page closed] | 17:03 | |
-!- gsomix [~gsomix@188.168.13.104] has quit [Quit: Leaving] | 17:41 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 268 seconds] | 18:18 | |
wiking | sonne|work, sonney2k around? | 18:20 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 18:47 | |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Ping timeout: 252 seconds] | 18:58 | |
-!- heiko1 [~heiko@host86-177-179-234.range86-177.btcentralplus.com] has left #shogun [] | 19:28 | |
@sonney2k | wiking, ? | 20:07 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has joined #shogun | 20:07 | |
-!- blackburn [~blackburn@188.168.13.104] has joined #shogun | 20:08 | |
wiking | sonney2k: i saw that yesterday you were talking about narray with heiko | 20:25 |
wiking | sonney2k: i'm missing a resizeable matrix in shogun currently | 20:26 |
wiking | so i was wondering if u guys came to a conclusion... | 20:26 |
@sonney2k | shogun-buildbot, force build nightly_default | 20:31 |
shogun-buildbot | build #151 forced | 20:31 |
shogun-buildbot | I'll give a shout when the build finishes | 20:31 |
@sonney2k | wiking, SGNDArray is what we talked about | 20:32 |
wiking | sonney2k: and do i recall it right that u were working with dna sequences? :) | 20:32 |
@sonney2k | wiking, an N-dimensional array | 20:32 |
@sonney2k | wiking, what is a resizable matrix for you? | 20:32 |
@sonney2k | wiking, yes but? | 20:33 |
wiking | sonney2k: well i currently have this in the new sosvm | 20:33 |
wiking | std::vector<SGSparseVector<float64_t> > | 20:33 |
wiking | i can resize both the vector and the inner sparsevector | 20:33 |
@sonney2k | wiking, why not SGSparseMatrix? | 20:33 |
wiking | without loosing the content of it | 20:33 |
@sonney2k | wiking, well you could use SGSparseMatrix. | 20:34 |
wiking | sonney2k: sgsparsematrix has no realloc option | 20:34 |
wiking | or i didn't see and i was blind | 20:34 |
wiking | no it doesn't have a line where it uses sg_realloc | 20:35 |
wiking | nor a function call for resizing | 20:35 |
@sonney2k | wiking, std::vector is no different | 20:35 |
wiking | sonney2k: i can resize the vector w/o problems | 20:35 |
@sonney2k | it is slow slow slow when resizing | 20:37 |
@sonney2k | you need to copy everything | 20:37 |
@sonney2k | why not allocate an empty SGSparseMatrix with whatever max number of vectors you need? | 20:37 |
@sonney2k | wiking, yeah but it same you could do with SGSparseMatrix | 20:37 |
wiking | http://www.cplusplus.com/reference/stl/vector/resize/ | 20:37 |
wiking | sonney2k: you mean to add a new function to it? :) | 20:37 |
wiking | and about the dna thing i was just looking for data set and thought you might have something or know somebody who would have something | 20:38 |
@sonney2k | wiking, yes sure - it is not exactly magic to do so. but I would rather resize in big chunks - but no idea if the resizing is time critical in your case | 20:38 |
@sonney2k | wiking, what exactly are you looking for? | 20:39 |
wiking | no it's it's not time critical that much... i mean there are other parts of that sosvm that are way slower so this wouldn't really slow down the thing | 20:39 |
wiking | sonney2k: motif finding | 20:39 |
@sonney2k | wiking, ok - and you really need sparse not SGStringList right? | 20:43 |
wiking | indeed | 20:43 |
wiking | the two things are not really related | 20:43 |
@sonney2k | wiking, no I mean instead of your sparse vector / matrix list | 20:44 |
wiking | well i only need to be | 20:47 |
wiking | to resize both horizontally and vertically the matrix | 20:48 |
wiking | without all the time maintaining | 20:49 |
wiking | the values | 20:49 |
@sonney2k | wiking, well SGStringList is the dense variant of SGSparseMatrix | 20:50 |
@sonney2k | wiking, actually a list of vectors | 20:51 |
wiking | yeah but it's the same deal | 20:52 |
wiking | i cannot resize this dynamically | 20:53 |
@sonney2k | wiking, well the same way | 20:53 |
@sonney2k | anyways you could do a couple of things but they all take quite some time | 20:54 |
shogun-buildbot | build #151 of nightly_default is complete: Success [build successful] Build details are at http://www.shogun-toolbox.org/buildbot/builders/nightly_default/builds/151 | 21:06 |
wiking | hehe http://imgur.com/t5Qvz | 21:10 |
wiking | :> | 21:10 |
wiking | blackburn: ^ | 21:11 |
blackburn | wiking: :) | 21:13 |
@sonney2k | hurray | 21:21 |
@sonney2k | http://shogun-toolbox.org/doc/en/latest/shogun_tutorial.pdf | 21:21 |
wiking | hahaha | 21:31 |
wiking | wtf is this turbo boost ? | 21:31 |
wiking | are we back to the 90s? | 21:31 |
-!- n4nd0 [~nando@s83-179-44-135.cust.tele2.se] has quit [Quit: leaving] | 21:36 | |
blackburn | wiking: turbo boost? | 21:48 |
wiking | blackburn: intel's new shit :) | 21:48 |
wiking | in the sandy/ivy-bridge arch | 21:48 |
-!- blackburn [~blackburn@188.168.13.104] has quit [Quit: Leaving.] | 22:27 | |
-!- blackburn [~blackburn@188.168.13.104] has joined #shogun | 22:31 | |
-!- ptizoom [~christian@85.210.80.20] has quit [Quit: Ex-Chat] | 23:26 | |
-!- ptizoom_ [~christian@85.210.80.20] has joined #shogun | 23:26 | |
--- Log closed Tue Oct 30 00:00:17 2012 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!