--- Log opened Fri Apr 08 00:00:36 2011 | ||
-!- blackburn [~qdrgsm@188.168.5.43] has left #shogun [] | 00:29 | |
-!- ameerkat [HydraIRC@149-169-253-238.nat.asu.edu] has quit [Ping timeout: 240 seconds] | 00:47 | |
-!- Pinio [~IceChat7@90-156-17-126.internetia.net.pl] has quit [Quit: IceChat - Its what Cool People use] | 00:48 | |
josip | what would be the best way to attach my transcripts/cv to the application? | 00:59 |
---|---|---|
josip | I can provide dropbox links or e-mail them to .. someone | 00:59 |
alesis-novik | your what? | 01:08 |
alesis-novik | I guess there is the field for additional info | 01:08 |
josip | but I don't feel very comfortable putting my transcripts online | 01:30 |
-!- Mengyun [~dorothyk@S0106001a803d4363.vc.shawcable.net] has joined #shogun | 01:38 | |
alesis-novik | Why do you want to put a transcript? | 01:58 |
alesis-novik | In cases like these I generally state what I did/have, and ask to be contacted if proof is required (but don't take my work for this being a good or bad idea, I'm just an applicant like you) | 02:00 |
alesis-novik | So what idea are you applying for? | 02:02 |
josip | the boosting-like algorithm for variable-dimensinaol features | 02:20 |
josip | the last one on the list if that rings a bell | 02:20 |
-!- epps [~epps@unaffiliated/epps] has quit [Read error: Operation timed out] | 02:51 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 04:47 | |
siddharth | sry sonney2k went to sleep after mailing you | 05:02 |
siddharth | will mail the patch on mailing list after some more changes | 05:02 |
-!- epps [~epps@unaffiliated/epps] has joined #shogun | 05:08 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun | 05:20 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 05:24 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 05:58 | |
siddharth | anyone here? | 06:06 |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 06:27 | |
serialhex | not really :P | 06:29 |
serialhex | bettyboo: whats up cutie? | 06:29 |
@bettyboo | serialhex: btw, you said there is about ten proposals? | 06:29 |
serialhex | bettyboo: no, many more than that | 06:29 |
@bettyboo | serialhex: I mean, there are only so many good ML researchers with that kinda background who could do all of those problems, and then the number of students would be like, maybe 20 times that? I can't imagine there being more than 400 of us | 06:30 |
serialhex | bettyboo: how many people propose to you? | 06:30 |
@bettyboo | serialhex: what algorithms do you propose to implement? | 06:30 |
@bettyboo | yes | 06:30 |
serialhex | bettyboo: fie on you! fie fie fie! | 06:30 |
@bettyboo | serialhex: nasty you!@ | 06:30 |
serialhex | :P | 06:30 |
@bettyboo | ;> | 06:30 |
serialhex | bettyboo: you're my kind of girl!! | 06:31 |
@bettyboo | serialhex: what was the train/test split, did you use 10cfv? how long did the training last, which kernel, what kind of feature data, binary, float, decimal, how long did the testing last ? | 06:31 |
serialhex | bettyboo: the training lasted a long time! | 06:31 |
@bettyboo | serialhex: are there any c# wrappers for it? the other, they report 10mil training samples,do you know how big was each sample (without going into those papers)? | 06:31 |
-!- siddharth [~siddharth@117.211.88.150] has quit [Remote host closed the connection] | 07:04 | |
-!- siddharth_ [~siddharth@117.211.88.150] has joined #shogun | 07:05 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 07:12 | |
-!- Mengyun [~dorothyk@S0106001a803d4363.vc.shawcable.net] has quit [Read error: Connection reset by peer] | 07:35 | |
-!- josip [~josip@unaffiliated/josip] has quit [Remote host closed the connection] | 07:42 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 07:58 | |
-!- lionelc [48cb93da@gateway/web/freenode/ip.72.203.147.218] has quit [Quit: Page closed] | 08:33 | |
-!- josip [~josip@unaffiliated/josip] has joined #shogun | 09:45 | |
josip | Hello, is there any mechanism in shogun to handle cases when a feature can belong to multiple classes? | 09:46 |
josip | so you can of course always recode the labels, but is there any class specifically designed for this? | 09:47 |
-!- siddharth_ [~siddharth@117.211.88.150] has quit [Ping timeout: 246 seconds] | 09:52 | |
@sonney2k | josip, no | 09:57 |
-!- siddharth_ [~siddharth@117.211.88.150] has joined #shogun | 10:01 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 10:18 | |
-!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 10:40 | |
-!- aifargonos [~aifargono@46.18.27.4] has joined #shogun | 10:40 | |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 260 seconds] | 10:50 | |
-!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 10:59 | |
-!- Ziyuan [~Ziyuan@116.22.39.134] has joined #shogun | 11:24 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 11:46 | |
josip | sonney2k: okay, thanks | 12:03 |
-!- skydiver [c255a037@gateway/web/freenode/ip.194.85.160.55] has joined #shogun | 12:08 | |
-!- eppz [~epps@unaffiliated/epps] has joined #shogun | 12:10 | |
-!- epps [~epps@unaffiliated/epps] has quit [Read error: Connection reset by peer] | 12:10 | |
-!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 12:20 | |
-!- aifargonos [~aifargono@46.18.27.4] has quit [Ping timeout: 240 seconds] | 12:34 | |
-!- siddharth_ [~siddharth@117.211.88.150] has quit [Read error: Connection reset by peer] | 12:36 | |
-!- siddharth [~siddharth@117.211.88.150] has joined #shogun | 12:38 | |
-!- Tanmoy [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 12:45 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 12:55 | |
-!- makokal [~nyumbanil@212.201.44.245] has joined #shogun | 13:13 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 13:16 | |
-!- blackburn [~qdrgsm@188.168.5.184] has joined #shogun | 13:40 | |
-!- aifargonos [~aifargono@46.18.27.4] has joined #shogun | 13:43 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 14:26 | |
-!- makokal [~nyumbanil@212.201.44.245] has quit [Ping timeout: 240 seconds] | 14:27 | |
-!- skydiver [c255a037@gateway/web/freenode/ip.194.85.160.55] has quit [Quit: Page closed] | 14:40 | |
-!- makokal [~nyumbanil@212.201.44.243] has joined #shogun | 14:48 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Ping timeout: 252 seconds] | 14:52 | |
siddharth | sonney2k, I have submitted the patch in the mailing list | 15:08 |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 15:35 | |
-!- eppz [~epps@unaffiliated/epps] has quit [Read error: Connection reset by peer] | 15:40 | |
-!- eppz [~epps@84.18.157.200] has joined #shogun | 15:40 | |
-!- eppz [~epps@84.18.157.200] has quit [Changing host] | 15:40 | |
-!- eppz [~epps@unaffiliated/epps] has joined #shogun | 15:40 | |
-!- oiwah [~oiwah@i114-181-177-143.s04.a013.ap.plala.or.jp] has joined #shogun | 16:04 | |
-!- oiwah [~oiwah@i114-181-177-143.s04.a013.ap.plala.or.jp] has quit [Client Quit] | 16:04 | |
-!- Chen [cab5d6cd@gateway/web/freenode/ip.202.181.214.205] has joined #shogun | 16:06 | |
Chen | Hello, I am student and I want to apply the GSoC project. I have successfully installed shogun on Ubuntu. So, I can submit the proposal now? | 16:08 |
-!- aifargonos [~aifargono@46.18.27.4] has quit [Ping timeout: 276 seconds] | 16:08 | |
-!- Chen [cab5d6cd@gateway/web/freenode/ip.202.181.214.205] has quit [Ping timeout: 252 seconds] | 16:19 | |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 16:31 | |
Ziyuan | Is it too late? | 16:34 |
makokal | @ziyuan too late for what? | 16:35 |
eppz | http://bit.ly/eYo7yf | 16:37 |
josip | is it really needed to write an e-mail to the mailing list that I managed to intall shogun? | 16:43 |
josip | why not just write that in the proposal | 16:43 |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 16:45 | |
-!- aifargonos [~aifargono@46.18.27.35] has joined #shogun | 16:53 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 17:07 | |
serialhex | yeah, you have a little more than 3 hours | 17:07 |
-!- Mengyun [~dorothyk@S0106001a803d4363.vc.shawcable.net] has joined #shogun | 17:07 | |
serialhex | ...what eppz said! :P | 17:08 |
@bettyboo | ;> | 17:08 |
-!- Mengyun [~dorothyk@S0106001a803d4363.vc.shawcable.net] has quit [Ping timeout: 260 seconds] | 17:13 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 17:22 | |
-!- josip [~josip@unaffiliated/josip] has quit [Remote host closed the connection] | 17:28 | |
* sonney2k is desperate | 17:29 | |
@sonney2k | over 60 students applied so far - there is now way to make a fair ranking - and we seriously cannot handly more than 10 of you :`-( | 17:31 |
siddharth | hi sonney2k | 17:33 |
@bettyboo | moin. siddharth | 17:33 |
siddharth | i have send the patch to the mailing list | 17:33 |
Ziyuan | Is src/libshogun the same as src/shogun ? | 17:34 |
siddharth | hi betty | 17:34 |
@bettyboo | siddharth: ask to put a C -> F conversion module into betty | 17:34 |
@bettyboo | ho siddharth | 17:34 |
siddharth | haha...those were sonney2k words :D | 17:34 |
@bettyboo | :> | 17:34 |
@sonney2k | Ziyuan, it is a symlink yes | 17:34 |
@sonney2k | siddharth, I am looking at your patch right now | 17:35 |
@sonney2k | looks pretty good! | 17:35 |
siddharth | ok :) | 17:35 |
@sonney2k | minor whitespace issues | 17:35 |
siddharth | yeah | 17:35 |
@sonney2k | you mixed tabs with spaces | 17:35 |
siddharth | i forgot to correct | 17:35 |
@sonney2k | and missing newlines between functions but that is really all | 17:35 |
@sonney2k | let me merge it | 17:35 |
Ziyuan | Ah, symlink, thanks | 17:36 |
siddharth | well I have done for matlab interface only | 17:36 |
@sonney2k | siddharth, actually it says siddharth <siddharth@siddharth-laptop.(none)> | 17:36 |
siddharth | will do for other too | 17:36 |
@sonney2k | do you want to give reasonable email addr's? | 17:36 |
siddharth | yes...new with git | 17:36 |
siddharth | yes | 17:37 |
siddharth | siddharthkherada27@gmail.com | 17:37 |
* sonney2k just set up http://shogun-toolbox.org/trac | 17:37 | |
@sonney2k | siddart you can use git config to set your username | 17:38 |
@sonney2k | e.g. I have | 17:38 |
@sonney2k | git config -l | 17:38 |
@sonney2k | user.name=Soeren Sonnenburg | 17:38 |
@sonney2k | user.email=sonne@debian.org | 17:38 |
siddharth | ok I will do that | 17:38 |
siddharth | thanks :) | 17:38 |
@bettyboo | ;D | 17:39 |
@sonney2k | siddharth: http://help.github.com/git-email-settings/ | 17:39 |
@sonney2k | siddharth, would you mind to send the patch again when you have done this | 17:39 |
@sonney2k | it is better that you type your name and email addr correctly | 17:40 |
@sonney2k | I cannot change it afterwards | 17:40 |
siddharth | sonney2k, can I send after the deadline? | 17:40 |
@sonney2k | siddharth, ahh you are still writing on your proposal? | 17:40 |
siddharth | yeah | 17:40 |
@sonney2k | siddharth, just send it quickly to me personally - it really is just issuing | 17:41 |
@sonney2k | $ git config --global user.name "Firstname Lastname"Sets the name of the user for all git instances on the system | 17:41 |
@sonney2k | $ git config --global user.email "your_email@youremail.com" | 17:41 |
@sonney2k | and git reformat-patch again | 17:41 |
@sonney2k | that's it | 17:41 |
siddharth | ok doing now | 17:41 |
serialhex | wow, yeah, just from the people who are saying they've set up shogun theres a bunch!! | 17:47 |
siddharth | sonney2k, should i send to you or the mailing list? | 17:51 |
@sonney2k | serialhex, I was seriously only expecting 10-15 people... | 17:52 |
@sonney2k | siddharth, to me is sufficient / but up to you | 17:52 |
@sonney2k | now I am already scared of having to wade through all of them | 17:52 |
serialhex | sonney2k: you have what? like 2 weeks to go through them?? you could turn around and ask Q's of the applicants can't you?? | 17:53 |
serialhex | ...kind of like a job interview? (i mean, this kind of is) | 17:53 |
siddharth | sonney2k, done :) | 17:54 |
@bettyboo | :> | 17:54 |
@sonney2k | serialhex, we have to reduce the number of potential candidates very quickly - it is totally impossible to interview >60 people | 17:54 |
serialhex | very true, very true | 17:54 |
@sonney2k | siddharth, thanks - very good. applying now | 17:54 |
siddharth | sonney2k, there are 60 proposals or 60 students? :O | 17:55 |
@sonney2k | siddharth, that also means I have to convince 10-15 potential mentors to go through all of them... | 17:55 |
serialhex | ...is there anything i can do to help my application? | 17:55 |
@sonney2k | serialhex, patches like siddharth sent... | 17:56 |
siddharth | sonney2k, should I mention patch in proposal? | 17:56 |
@sonney2k | siddharth, of course!!! | 17:56 |
brocolli | heh | 17:56 |
serialhex | sonney2k: ok, i'll see what i can find to do... | 17:56 |
siddharth | ok :) | 17:56 |
brocolli | I am going to be laughing if I get rejected because my name on irc isn't proper :/ | 17:56 |
brocolli | or somethign like tha | 17:56 |
Ziyuan | In center_matrix of Mathematics.h, matrix[int64_t(i)*m+j]+=s-colsums[i]-rowsums[j]; should be matrix[int64_t(i)*m+j]+=s-colsums[j]-rowsums[i]; | 17:56 |
* serialhex looks for something a n00b can code :P | 17:57 | |
Ziyuan | Right? | 17:57 |
@sonney2k | serialhex, it is really >60 students and even more proposals... - it is crazy | 18:05 |
@sonney2k | makes you proud and ashamed at the same time :( | 18:05 |
brocolli | what | 18:06 |
brocolli | how the hell | 18:06 |
Ziyuan | Let me make a patch | 18:06 |
brocolli | I don't believe there are that many machine learners out there sonney2k | 18:06 |
brocolli | :/ | 18:06 |
@sonney2k | brocolli, me neither - so lets hope half of them are not and stand corrected !? | 18:07 |
brocolli | lol | 18:07 |
brocolli | I mean, the only reaon I applied was to see if I got it + needed a job. Otherwise if I'm getting funded over the summer I'd rather work on other stuff ;p | 18:07 |
@sonney2k | Ziyuan, could be - the matrix is supposedly square so I never checked | 18:08 |
brocolli | But I don't know if I will be funded yet, hence my application ;) | 18:08 |
brocolli | It's admirable people want to do ML, but it's more than just programming. GSoC == Funding but simultaneously encourages people to apply even if it's not their area of interst ;P | 18:09 |
CIA-31 | shogun: Siddharth Kherada master * r9f7760c / (6 files in 2 dirs): Wavelet Kernel Function - http://bit.ly/i846Sl | 18:10 |
CIA-31 | shogun: Soeren Sonnenburg master * r4ae3a56 / (2 files): Minor reformatting, inlining and whitespace changes - http://bit.ly/fT7QST | 18:11 |
CIA-31 | shogun: Soeren Sonnenburg master * rd52c6de / src/modular/Kernel.i : Add modular interface to Wavelet Kernel - http://bit.ly/hQxeAx | 18:11 |
@sonney2k | siddharth, patch applied - thanks, you should also add some GPL copyright header or I will do (sufficient to do this later) | 18:12 |
@sonney2k | brocolli, err so what shall we do now, not select you? When will you know? | 18:13 |
brocolli | Well I should know in a day or two | 18:13 |
brocolli | I mean once you guys have made a decision on me, before making it official can you let meknow | 18:14 |
brocolli | I'll let you know the same day | 18:14 |
siddharth | sonney2k, ok i will do after the deadline | 18:15 |
@sonney2k | brocolli, I guess we only start looking on monday | 18:16 |
-!- siddharth [~siddharth@117.211.88.150] has quit [Remote host closed the connection] | 18:24 | |
eppz | sonney2k, since you have so many people interested, you could organize a mentoring project outside the GSoC framework | 18:26 |
-!- siddharth [~siddharth@117.211.88.150] has joined #shogun | 18:26 | |
brocolli | no one gets funded that way | 18:27 |
brocolli | the only reason students get mentored is because they're paid | 18:27 |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has joined #shogun | 18:27 | |
eppz | not all | 18:28 |
eppz | Apache has a system like this | 18:28 |
brocolli | Everyone who is in GSoC gets paid. | 18:29 |
brocolli | the students at least. | 18:29 |
eppz | I know, but some students might also be in it for the learning experience and open source love. | 18:30 |
-!- argetlahmsource [~argetlahm@adsl-99-160-142-89.dsl.euclwi.sbcglobal.net] has joined #shogun | 18:31 | |
brocolli | heh, mentors don't have time for that, these are people who are active researchers man. | 18:31 |
eppz | they don't get paid for GSoC either | 18:32 |
eppz | I think the organization receives just $500 | 18:32 |
brocolli | But hte software is representing the institution | 18:33 |
brocolli | and in turn this is like a mentor's duty to his institution type thing | 18:34 |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has joined #shogun | 18:34 | |
eppz | not everything revolves on money, the open source movement is proof of that | 18:35 |
brocolli | lol | 18:35 |
eppz | what? | 18:36 |
brocolli | You're missing the point man, that's all. | 18:36 |
brocolli | There are plenty of open source projects that you can contribute to, including this one I'm sure. But the mentor part isn't going to happen I don't think, it might. But they're usually busy | 18:37 |
Ziyuan | Well...I added the file and committed it in git. Then I found my email not correct so I made a reset. After I rewrite my email and git commit again, it told me that nothing to commit. Now how I can get things back to normal? | 18:37 |
eppz | meh.. | 18:38 |
siddharth | Ziyuan, I think the changes will appear after next commit... | 18:47 |
-!- makokal [~nyumbanil@212.201.44.243] has left #shogun [] | 18:47 | |
Ziyuan | Ok, then mail the patch to sonne@debian.org ? | 18:55 |
Ziyuan | Done | 18:57 |
blackburn | sonney2k: guten Abend, another patch is in ML | 18:58 |
* blackburn wonders about number of proposals and students told by Sören | 19:03 | |
-!- eppz [~epps@unaffiliated/epps] has quit [Read error: Connection reset by peer] | 19:11 | |
-!- epps [~epps@84.18.157.200] has joined #shogun | 19:12 | |
-!- epps [~epps@84.18.157.200] has quit [Changing host] | 19:12 | |
-!- epps [~epps@unaffiliated/epps] has joined #shogun | 19:12 | |
-!- josip [~josip@212.201.44.246] has joined #shogun | 19:20 | |
-!- josip [~josip@212.201.44.246] has quit [Changing host] | 19:20 | |
-!- josip [~josip@unaffiliated/josip] has joined #shogun | 19:20 | |
josip | deadline -0140 | 19:20 |
josip | hehe | 19:20 |
@bettyboo | hihi!! | 19:20 |
josip | the melange editor is awful btw | 19:22 |
Ziyuan | Can I update my proposal after the deadline? | 19:24 |
josip | no | 19:30 |
@sonney2k | brocolli, we are surely not doing this for money (actually we don't get any) | 19:38 |
@sonney2k | for that | 19:38 |
brocolli | not you | 19:39 |
@sonney2k | I do it to get new contributors and be assured that I have to do all the real work while I am trying to mentor | 19:39 |
@sonney2k | well none of the people in the shogun project | 19:39 |
brocolli | Your institution is doing it because of the cutting edge nature of the project | 19:39 |
brocolli | it's implementing really good algorithms into an open source piece of software, all on behalf of MPI | 19:39 |
brocolli | It's good for the brand of the institution, I'm sure you cannot dispute that. | 19:39 |
josip | it's a win-win situation, that's why they're doing it | 19:39 |
@sonney2k | brocolli, the institution does not care - to the contrary they would rather want to see me do some real work (aka research) | 19:40 |
josip | you're working at MPI Tuebingen? | 19:40 |
brocolli | sonney2k: so what is the purpose of such a large undertaking? For fun | 19:40 |
brocolli | I say this with the utmost respect | 19:41 |
brocolli | I am just saying there are always hidden variables with large weight :P | 19:41 |
@bettyboo | yeah | 19:41 |
@sonney2k | brocolli, no idea.... look I am one of the initiators and still leading figure behind mloss.org / mldata.org | 19:42 |
* brocolli doesn't know much about open source but has heard of mloss.org for sure | 19:42 | |
@sonney2k | brocolli, well I wrote most of the website | 19:43 |
josip | thanks for mloss! It has proven very helpful to me | 19:43 |
@sonney2k | have a look at mldata.org too | 19:45 |
josip | yu; | 19:45 |
josip | p | 19:45 |
@sonney2k | mldata.org is like 100 times more work :D | 19:45 |
@bettyboo | yep | 19:45 |
@sonney2k | anyway got to go | 19:45 |
@sonney2k | cu | 19:45 |
@bettyboo | cya | 19:45 |
josip | I participated in last year's gsoc as well, that's not a problem - right? The FAQ says it's good | 19:45 |
CIA-31 | shogun: Ziyuan Lin master * r0723320 / src/libshogun/lib/Mathematics.h : Correct a subscript error - http://bit.ly/fa3JK4 | 19:47 |
Ziyuan | Just for curiosity, what project were you in last year? | 19:48 |
josip | python software foudnation, distutils2 | 19:48 |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 19:49 | |
Ziyuan | wow | 19:50 |
josip | ? | 19:51 |
-!- Ziyuan [~Ziyuan@116.22.39.134] has quit [] | 20:16 | |
-!- Ziyuan [~Ziyuan@116.22.39.134] has joined #shogun | 20:23 | |
-!- Ziyuan [~Ziyuan@116.22.39.134] has quit [Client Quit] | 20:24 | |
-!- Ziyuan [~Ziyuan@116.22.39.134] has joined #shogun | 21:03 | |
josip | good luck everyone! | 21:03 |
-!- alesis-novik [~alesis@188.74.87.84] has quit [Quit: I'll be Bach] | 21:05 | |
-!- argetlahmsource [~argetlahm@adsl-99-160-142-89.dsl.euclwi.sbcglobal.net] has left #shogun [] | 21:42 | |
@sonney2k | ok, so we have exactly 70 proposals now | 21:59 |
@sonney2k | by over 60 students | 22:00 |
epps | congrats | 22:00 |
@sonney2k | in #gsoc they say 417 mentoring orgs applied and 175 were accepted this year; 5474 proposals were submitted by 3731 students. | 22:00 |
@sonney2k | which means on average an organization gets 20 students / or 31 applications | 22:01 |
blackburn | really wonder how you will choose 5-10 from that 70 | 22:01 |
* sonney2k does not know either :-( | 22:01 | |
josip | sonney2k: I don't think the distribution is inoform :D | 22:01 |
@bettyboo | ;D | 22:01 |
@sonney2k | surely not | 22:01 |
josip | e.g. there are a lot of people applying for wordpress and stuff liek that | 22:01 |
blackburn | machine learning is in trending now | 22:01 |
@sonney2k | looks like | 22:01 |
blackburn | by some reason | 22:02 |
josip | well, it started being widely applied? | 22:02 |
josip | to almost everything? :) | 22:02 |
@sonney2k | the problem is we are a new organization - it is not clear how much google will trust us to mentor more than 5 students | 22:02 |
@bettyboo | yep :> | 22:02 |
@sonney2k | (the gsoc orga faq says 5 students for new orgas at the max) | 22:03 |
@sonney2k | anyways, we will do our best to get more... | 22:03 |
josip | hehe, good luck with that :) | 22:03 |
blackburn | sonney2k: 5 from 70? oh, I don't envy you | 22:03 |
@sonney2k | lets hope for the best | 22:05 |
-!- Mengyun [~Mengyun@S0106001a803d4363.vc.shawcable.net] has joined #shogun | 22:06 | |
blackburn | sonney2k: when creating new kernel, where should I describe it (excluding its .cpp and .h and kernel type in Kernel.h)? | 22:07 |
@sonney2k | well only in the doxygen doc in the .h and make an example | 22:08 |
blackburn | I mean I created it and it doesn't importing | 22:08 |
blackburn | okay, so think I did something wrong :) | 22:09 |
siddharth | did you do make install? | 22:09 |
blackburn | of course | 22:09 |
siddharth | no error in that? | 22:09 |
blackburn | nope | 22:09 |
blackburn | oh | 22:09 |
blackburn | I forget configure! | 22:09 |
@sonney2k | blackburn, you mean for the modular interfaces or what? | 22:09 |
siddharth | not needed | 22:09 |
blackburn | sonney2k: yeap, python_modular | 22:10 |
blackburn | sonney2k: siddharth's kernel is importing but my - not | 22:10 |
siddharth | did u change in GUIkernel file? | 22:10 |
@sonney2k | blackburn, open up modular/Kernel.i | 22:10 |
blackburn | sonney2k: ah! that's it | 22:10 |
blackburn | thank you | 22:10 |
@sonney2k | then search for GaussianKernel and add the same 3 lines :) | 22:11 |
blackburn | sonney2k: btw, my _proper_ patch is in ML | 22:11 |
blackburn | I was terrified when looked what I sent to ML | 22:12 |
blackburn | some trash generated by my eclipse, sorry :) | 22:12 |
@bettyboo | ;D blackburn | 22:12 |
@sonney2k | blackburn, don't worry. Always remember the ML'ers can't code so whatever we send looks great to 'them' | 22:12 |
@sonney2k | ;-) | 22:12 |
@bettyboo | funny | 22:12 |
-!- akhil_ [75d35896@gateway/web/freenode/ip.117.211.88.150] has quit [Quit: Page closed] | 22:13 | |
blackburn | I thought that Kernel.i is filled automatically.. | 22:14 |
blackburn | Was I wrong? | 22:14 |
@sonney2k | blackburn, that is not done automatically | 22:14 |
blackburn | I see now | 22:14 |
siddharth | for static interfaces you have to change in SGInterface and GUIKErnel file | 22:15 |
@sonney2k | exactly. siddharth knows what he is talking about | 22:15 |
blackburn | sonney2k: I should do this too, right? | 22:16 |
* sonney2k hates the static interfaces though | 22:16 | |
@sonney2k | but we started with them... | 22:16 |
* siddharth loves matlab :D | 22:16 | |
blackburn | >>> from shogun.Kernel import WaveKernel | 22:16 |
blackburn | >>> wk = WaveKernel() | 22:16 |
blackburn | thank you! | 22:16 |
blackburn | works now | 22:16 |
siddharth | nice | 22:16 |
blackburn | will now write this stuff in SGInterface and etc | 22:17 |
-!- ameerkat [~ameerkat@184-98-140-155.phnx.qwest.net] has quit [Ping timeout: 240 seconds] | 22:18 | |
@sonney2k | blackburn, I think it would be a good idea that you send your kernel overview html page to the mailinglist - in case others would like to contribute a patch | 22:18 |
blackburn | sorry, don't understand you properly, should I describe this kernel in html? | 22:19 |
@sonney2k | I think it is the only way to reduce the number of applicants requiring some contribution ... | 22:19 |
siddharth | sonney2k, me too? | 22:20 |
@sonney2k | well you both already contributed | 22:20 |
blackburn | i'm looking forward for further contributions :) | 22:20 |
@bettyboo | lol! | 22:20 |
siddharth | same here :) | 22:20 |
@bettyboo | ;D | 22:21 |
blackburn | so, when introducing new kernel I should send it with examples and description, right? | 22:21 |
@sonney2k | blackburn, that would be perfect | 22:21 |
blackburn | is it necessary to do it in html? | 22:21 |
blackburn | I mean there is only one equation and nothing to describe detailly | 22:21 |
* siddharth off to sleep...gn | 22:22 | |
blackburn | sonney2k: anyway, I have no problem with doing this, may I ask you later about some details of it by email or in irc? | 22:23 |
@sonney2k | blackburn, not html doxygen! | 22:23 |
@sonney2k | look at gaussiankernel | 22:23 |
blackburn | ah | 22:23 |
blackburn | that's what you mean | 22:24 |
blackburn | okay, I understand now | 22:24 |
blackburn | do you want to reduce applicants-contributors? | 22:24 |
blackburn | number of* I mean | 22:25 |
@sonney2k | blackburn, please send the url of that paper to the mailinglist ... I will send an email around saying that we have so many submission and some of you asked how they could improve scores... by contributing a patch is my answer | 22:25 |
blackburn | sorry, don't understand you again. what is the paper I should url? | 22:26 |
blackburn | may be I'm tired a bit, sorry :) | 22:27 |
@sonney2k | blackburn, the url with the kernels | 22:30 |
@sonney2k | that is really an easy start learning to extend shogun | 22:30 |
blackburn | http://crsouza.blogspot.com/2010/03/kernel-functions-for-machine-learning.html#wave | 22:30 |
blackburn | this? | 22:30 |
blackburn | with no '#wave' | 22:30 |
@sonney2k | yes | 22:31 |
blackburn | sonney2k: you will have a terrible count of commits :) | 22:31 |
@sonney2k | blackburn, or maybe only from the 10 students we can then later on accept | 22:31 |
@sonney2k | :) | 22:31 |
@bettyboo | ;D | 22:31 |
blackburn | what should you do if ~10 will commit you all the time.. :) | 22:32 |
blackburn | sonney2k: will you go to sleep in nearest a half of hour? | 22:32 |
blackburn | sonney2k: I could send a patch with WaveKernel with this link | 22:32 |
@sonney2k | blackburn, just sent the patch in one email and then later on reply to my email with this link | 22:33 |
@sonney2k | I am jsut writing it | 22:33 |
@sonney2k | :) | 22:33 |
blackburn | but why you want to me send this url? :) | 22:34 |
blackburn | me to send* (oh, I have bad teachers of english) | 22:34 |
blackburn | sonney2k: and again, what's with generalized HIK patch? did I done all the things right? | 22:35 |
@sonney2k | blackburn, looked good but I need to have a closer look. | 22:35 |
@sonney2k | will apply it probably today still. | 22:35 |
blackburn | I have tested it this time :) | 22:36 |
@bettyboo | rahaha | 22:36 |
blackburn | that non-proper patch is pretty ugly diff of something I don't understand | 22:36 |
blackburn | understand why you get angry with it :) | 22:36 |
@bettyboo | blackburn, rahaha | 22:36 |
@sonney2k | blackburn, you should really use github infrastructure | 22:37 |
@sonney2k | I mean fork off shogun from git | 22:37 |
@bettyboo | nite | 22:37 |
@sonney2k | then modify your tree | 22:37 |
@sonney2k | then submit a pull request | 22:37 |
@sonney2k | that is coolest - but of course patches to the mailinglist are as welcome! | 22:37 |
blackburn | okay, I will grasp with it | 22:38 |
blackburn | it's because I'm novice in git | 22:38 |
blackburn | that static interfaces in libshogunui.. I don't like it much :) | 22:40 |
@sonney2k | blackburn, I hate them too | 22:40 |
@sonney2k | they need a cleanup | 22:40 |
@sonney2k | swig is soo much more elegant | 22:40 |
blackburn | every kernel need it, right? | 22:40 |
@sonney2k | in particular you touch the modular/*.i file once and you get interfaces to alll the languages. | 22:41 |
blackburn | I mean should I do this? | 22:41 |
blackburn | now* | 22:41 |
blackburn | seems so :) | 22:41 |
@sonney2k | you don't have to do static :) | 22:42 |
blackburn | I don't?? | 22:42 |
* sonney2k only uses the modular interfaces... | 22:42 | |
@sonney2k | anyways | 22:42 |
blackburn | may I sent a patch with only modular available? | 22:42 |
@sonney2k | blackburn, sure | 22:42 |
blackburn | okay | 22:42 |
@sonney2k | ok, I've sent the email to the mailinglist | 22:42 |
@sonney2k | I suggest you reply with this url - it is a good thing for students to pick one kernel and to implement it.... | 22:43 |
blackburn | should I answer you with that link? | 22:43 |
@sonney2k | yes | 22:43 |
blackburn | okay | 22:43 |
@sonney2k | the patch as a separate email please | 22:43 |
blackburn | np | 22:43 |
@sonney2k | thanks | 22:43 |
blackburn | hope will not sink myself with that suggestion :) | 22:43 |
@sonney2k | I have to leave now... | 22:44 |
@sonney2k | will be back later | 22:44 |
blackburn | okay | 22:44 |
-!- hiteshk [~hitesh_na@117.211.88.150] has joined #shogun | 22:44 | |
-!- epps [~epps@unaffiliated/epps] has quit [Ping timeout: 276 seconds] | 22:46 | |
-!- hiteshk [~hitesh_na@117.211.88.150] has left #shogun [] | 22:47 | |
Mengyun | Hi sonney2k, is there any bug list for shogun available ? | 22:54 |
@sonney2k | Mengyun, not really an up-to-date one - I've just set up http://shogun-toolbox.org/trac - and tried to free it from some spam | 23:16 |
blackburn | sonney2k: I have introduced a new kernel, set up to modular, etc | 23:19 |
blackburn | should patch include doxygen docs for it? | 23:19 |
@sonney2k | blackburn, yes please | 23:19 |
blackburn | I mean doxygen generates it by source | 23:19 |
@sonney2k | it is not so difficult | 23:19 |
@sonney2k | just look at the comments in gaussiankernel.h for example | 23:19 |
@sonney2k | (only the .h file needs to be documented) | 23:20 |
blackburn | i done all this way | 23:20 |
blackburn | have done | 23:20 |
blackburn | okay, will create docs for it | 23:20 |
@sonney2k | blackburn, the doc will be autogenerated | 23:22 |
blackburn | I mean it | 23:22 |
@sonney2k | you can try to import your kernel form python and then do help(yourkernel) | 23:22 |
@sonney2k | it should show the docs | 23:22 |
@sonney2k | and the html site | 23:22 |
blackburn | just need to reconfig, disabled docs with slow generation :) | 23:22 |
@bettyboo | ;> | 23:23 |
blackburn | that kernel is a bit strange.. will cause many NaN when K(x,y) and x==y | 23:24 |
@sonney2k | doesn't sound right - maybe only for certain parameter settings? | 23:26 |
blackburn | there is theta/||x-y|| | 23:27 |
blackburn | x==y will cause theta/0 | 23:27 |
@sonney2k | you have to check if that is the right thing | 23:27 |
blackburn | I could better implement Cauchy kernel with no similar problems :D | 23:28 |
blackburn | http://latex.codecogs.com/png.latex?k(x,%20y)%20=%20\frac{\theta}{\lVert%20x-y%20\rVert%20\right}%20\sin%20\frac{\lVert%20x-y%20\rVert%20}{\theta} | 23:29 |
blackburn | that is wave kernel from link, I posted to ML | 23:29 |
@sonney2k | blackburn, I suspect that when ||x-y|| == 0 that the kernel == 0 | 23:29 |
@sonney2k | or 1 | 23:29 |
@sonney2k | you have to check | 23:29 |
blackburn | there is a link on that page to some dissertation.. and there is only mention of it | 23:31 |
blackburn | with no Mercer prove or something else | 23:31 |
blackburn | seems to be a very strange kernel | 23:31 |
Mengyun | thanks sonney2k | 23:31 |
blackburn | Limit[(h/x)*Sin[x/h], x -> 0] | 23:32 |
blackburn | 1 | 23:32 |
@sonney2k | blackburn, I think you could just replace sin(x) by its taylor expansion and then you will hopefully see what happens when ||x-y|| -> 0 | 23:32 |
blackburn | I shouldn't could use that limit ^^ | 23:32 |
blackburn | seems that 1 | 23:33 |
@sonney2k | blackburn, I would expect that the value is 1.0 when x==y | 23:33 |
blackburn | sonney2k: yeap, and with no respect to theta it is 1.0 | 23:33 |
blackburn | if (dist==0.0) return 1.0; | 23:34 |
blackburn | dirty :) | 23:34 |
@sonney2k | blackburn, so just add a if statement and all good | 23:34 |
blackburn | thank you | 23:35 |
@sonney2k | Mengyun, what for? | 23:35 |
blackburn | sonney2k: trac | 23:35 |
@sonney2k | ahh yes | 23:36 |
Mengyun | yep | 23:36 |
* sonney2k looses trac(k) of everything | 23:36 | |
Mengyun | :] | 23:36 |
@sonney2k | Mengyun, the GSoC ideas in trac are from 2010 - we got rejected back then... | 23:37 |
@bettyboo | hiho!! | 23:37 |
* sonney2k what is going on with betty | 23:38 | |
blackburn | sonney2k: | 23:39 |
blackburn | Help on class WaveKernel in module shogun.Kernel: | 23:39 |
blackburn | class WaveKernel(BaseKernel) | 23:39 |
blackburn | | Wave kernel, formally described as \[ K(x,x') = | 23:39 |
blackburn | | \frac{\theta}{\| x-y \|} \sin \frac{\| x-y | 23:39 |
blackburn | | \|}{\theta} \]. | 23:39 |
blackburn | | | 23:39 |
blackburn | right? | 23:39 |
blackburn | it is in python I mean :) | 23:40 |
@sonney2k | yes | 23:40 |
@sonney2k | the generated html code will look better | 23:40 |
blackburn | I saw it | 23:40 |
@sonney2k | I mean it is latex compiled | 23:40 |
blackburn | should I send xml? | 23:40 |
blackburn | oops | 23:41 |
blackburn | not xml | 23:41 |
@sonney2k | blackburn, ehh no, just the source code - the doxygen doc can be automatically generated | 23:41 |
blackburn | should it be in doc? | 23:41 |
@sonney2k | IIRC doc/html | 23:42 |
blackburn | I see now | 23:43 |
blackburn | seems that I should make it too | 23:43 |
blackburn | 'make' | 23:43 |
-!- blackburn1 [~qdrgsm@188.168.4.21] has joined #shogun | 23:47 | |
-!- blackburn [~qdrgsm@188.168.5.184] has quit [Ping timeout: 258 seconds] | 23:48 | |
blackburn1 | sorry, disconnected | 23:48 |
blackburn1 | for some reason there is not that equation in .html .. | 23:51 |
@sonney2k | blackburn1, it is not in html but some .png IIRC | 23:55 |
blackburn1 | yeap, not that .png is in .html | 23:55 |
blackburn1 | btw, have this thing: | 23:55 |
blackburn1 | ! LaTeX Error: Illegal character in array arg. | 23:55 |
blackburn1 | See the LaTeX manual or LaTeX Companion for explanation. | 23:55 |
blackburn1 | Type H <return> for immediate help. | 23:55 |
blackburn1 | ... | 23:55 |
blackburn1 | 23:55 | |
blackburn1 | l.17 \[ \begin{array}( | 23:55 |
blackburn1 | {ccccc}) 0 & \alpha^{(1)} & \alpha^{(2)} & \cdots & \a... | 23:55 |
blackburn1 | ({ccccc}) isn't a proper way to describe array in latex | 23:56 |
-!- blackburn1 was kicked from #shogun by bettyboo [flood] | 23:56 | |
-!- blackburn1 [~qdrgsm@188.168.4.21] has joined #shogun | 23:56 | |
blackburn1 | sorry :) | 23:56 |
blackburn1 | sonney2k: should I find there is that error? | 23:57 |
blackburn1 | I mean it isn't my fault, described somewhere else | 23:58 |
@sonney2k | blackburn1, it is in adaboost | 23:59 |
@sonney2k | hmmhh maybe I should remove the experimental boosting stuff for now | 23:59 |
--- Log closed Sat Apr 09 00:00:36 2011 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!