--- Log opened Sat Sep 06 00:00:30 2014 | ||
-!- PirosB3 [~pirosb3@host132-36-dynamic.9-87-r.retail.telecomitalia.it] has joined #shogun | 01:49 | |
-!- PirosB3 [~pirosb3@host132-36-dynamic.9-87-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 03:33 | |
shogun-buildbot | build #901 of nightly_default is complete: Failure [failed notebooks] Build details are at http://buildbot.shogun-toolbox.org/builders/nightly_default/builds/901 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com>, khalednasr <k.nasr92@gmail.com> | 03:55 |
---|---|---|
-!- PirosB3 [~pirosb3@host132-36-dynamic.9-87-r.retail.telecomitalia.it] has joined #shogun | 12:24 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 15:23 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has quit [Remote host closed the connection] | 15:28 | |
-!- abinash_panda [~abinash@116.75.97.43] has joined #shogun | 19:53 | |
-!- kislay [~Abhijeet@117.239.94.112] has joined #shogun | 19:57 | |
-!- jiaolong [58a0e807@gateway/web/freenode/ip.88.160.232.7] has joined #shogun | 20:09 | |
-!- lambday [7aa77165@gateway/web/freenode/ip.122.167.113.101] has joined #shogun | 20:19 | |
-!- HeikoS [~heiko@dab-ntm1-h-1-2.dab.02.net] has joined #shogun | 20:39 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:39 | |
kislay | hi HeikoS | 20:44 |
-!- yorkerlin [18d4da3a@gateway/web/freenode/ip.24.212.218.58] has joined #shogun | 20:53 | |
-!- sonne|osx [~sonne@f052229200.adsl.alicedsl.de] has joined #shogun | 20:56 | |
@HeikoS | kislay: hey there! | 20:58 |
kislay | HeikoS, Kevin won't be able to make it. | 20:58 |
@HeikoS | kislay: ok | 20:58 |
@HeikoS | kislay: well you are here and can represent him | 20:59 |
kislay | yeah sure | 20:59 |
@HeikoS | ok let me check who is here | 20:59 |
@HeikoS | or who is missing | 20:59 |
-!- khalednasr [~k.nasr92@41.69.136.192] has joined #shogun | 21:00 | |
@HeikoS | abinash_panda, jiaolong, khalednasr, lambday, yorkerlin hi | 21:00 |
lambday | hi HeikoS | 21:00 |
jiaolong | hello | 21:00 |
khalednasr | hi | 21:00 |
yorkerlin | hi | 21:00 |
@HeikoS | khalednasr: theo coming? | 21:00 |
khalednasr | HeikoS, dunno | 21:00 |
@HeikoS | khalednasr: probably not, I met him and now know how unreliable his emails are ;) | 21:00 |
@HeikoS | lisitsyn, wiking hi# | 21:01 |
-!- pickle27 [~AndChat79@24.114.82.188] has joined #shogun | 21:03 | |
pickle27 | Hey guys sorry I can't stay long, but congrats everyone on gsoc! | 21:04 |
-!- thoralf [~thoralf@91.65.138.205] has joined #shogun | 21:05 | |
-!- mode/#shogun [+o thoralf] by ChanServ | 21:05 | |
@thoralf | Hello :) | 21:05 |
kislay | pickle27, yeah :) | 21:05 |
kislay | thoralf, Hi T! | 21:05 |
@thoralf | Hey K! | 21:06 |
-!- HeikoS [~heiko@dab-ntm1-h-1-2.dab.02.net] has quit [Ping timeout: 245 seconds] | 21:06 | |
@lisitsyn | hey | 21:07 |
@thoralf | Hey Sergey | 21:07 |
@thoralf | lisitsyn: How's your new job? | 21:08 |
-!- Saurabh7 [~guestie_@103.225.100.51] has joined #shogun | 21:08 | |
@lisitsyn | thoralf: pretty good! | 21:08 |
@lisitsyn | thoralf: though taking too much time :D | 21:08 |
@thoralf | lisitsyn: I can imagine (same here ;)) | 21:08 |
-!- kislay__ [~Abhijeet@117.239.94.112] has joined #shogun | 21:10 | |
@lisitsyn | thoralf: so heiko should be there to lead this meeting thing? :D | 21:11 |
-!- kislay [~Abhijeet@117.239.94.112] has quit [Ping timeout: 276 seconds] | 21:12 | |
lambday | lisitsyn: you've got a new job? | 21:14 |
@lisitsyn | lambday: yes | 21:15 |
lambday | lisitsyn: which company? | 21:15 |
@lisitsyn | lambday: yandex | 21:15 |
@lisitsyn | lambday: not only job but a new city :D | 21:15 |
lambday | lisitsyn: you're not in samara anymore | 21:15 |
lambday | ? | 21:15 |
@lisitsyn | lambday: now moscow | 21:16 |
lambday | awesome!! | 21:16 |
@lisitsyn | lambday: it was a good time to move on ;) | 21:16 |
lambday | lisitsyn: lol yeah getting old, eh? | 21:16 |
@lisitsyn | haha yeah | 21:16 |
@lisitsyn | so it is better to move before getting way too old | 21:16 |
@lisitsyn | :D | 21:16 |
lambday | lisitsyn: you're still way younger and I'm jealous :D | 21:17 |
@lisitsyn | lambday: if you don't know about yandex in simple words it's just a russian google :D | 21:17 |
lambday | lisitsyn: really!! frickin' awesome! | 21:17 |
pickle27 | Nice dude that sounds awesome | 21:17 |
lambday | lisitsyn: google people called me for a job offer but I still don't have my frickin' degree! :( | 21:17 |
@lisitsyn | lambday: wow cool, is it like a real offer? | 21:18 |
lambday | lisitsyn: not offer - I've gotta appear for an interview | 21:18 |
@lisitsyn | ahh | 21:18 |
@lisitsyn | I went through this interviw hell :D | 21:18 |
lambday | lisitsyn: sucks! | 21:18 |
lambday | lisitsyn: will try for Amazon! they already called for 6-7 times.. I'm hating it at Oracle | 21:19 |
@lisitsyn | lambday: ahh btw how is oracle? | 21:19 |
lambday | lisitsyn: company is good - my project - sucks! | 21:19 |
lambday | don | 21:19 |
lambday | don't want to work on such an application layer.. | 21:19 |
@lisitsyn | lambday: what do you do? | 21:20 |
lambday | mostly j2ee and cloud stuf | 21:20 |
-!- Heiko_web [05439144@gateway/web/freenode/ip.5.67.145.68] has joined #shogun | 21:20 | |
Heiko_web | hello? | 21:20 |
lambday | hi Heiko_web | 21:20 |
Heiko_web | I had some connectivity issues | 21:20 |
Heiko_web | very sorry | 21:20 |
Heiko_web | My client basically told me nobody was here :( | 21:20 |
@lisitsyn | haha | 21:20 |
Heiko_web | Who is around? | 21:20 |
lambday | I am | 21:21 |
Heiko_web | was just writing an angry email, haha | 21:21 |
Heiko_web | could we get a ping from everyone? | 21:21 |
lambday | Heiko_web: lol :P | 21:21 |
kislay__ | I am here! | 21:21 |
Heiko_web | just found out over the channel logs | 21:21 |
-!- HeikoS [~heiko@05439144.skybroadband.com] has joined #shogun | 21:21 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 21:21 | |
@HeikoS | test | 21:21 |
@HeikoS | ah now things work again, weird pidgin | 21:21 |
@lisitsyn | I am here | 21:21 |
-!- Heiko_web [05439144@gateway/web/freenode/ip.5.67.145.68] has quit [Client Quit] | 21:22 | |
@HeikoS | ok | 21:22 |
@thoralf | Hello H | 21:22 |
@HeikoS | so lets see pickle27, lisitsyn, thoralf wiking hi | 21:22 |
@HeikoS | kislay__: abinash_panda, jiaolong, khalednasr, Saurabh7, lambday, yorkerlin hi! | 21:22 |
kislay__ | hi HeikoS | 21:23 |
khalednasr | hi | 21:23 |
yorkerlin | hi again | 21:23 |
@HeikoS | very sorry for the delay and my previous issues | 21:23 |
Saurabh7 | oh, heelllo | 21:23 |
@HeikoS | I think the meeting can start | 21:23 |
@HeikoS | Saurabh7: hi nice to see you | 21:23 |
@HeikoS | yorkerlin: hi! emt will not make it | 21:23 |
@HeikoS | lambday: dino probably also not | 21:23 |
@HeikoS | but thats fine | 21:23 |
@HeikoS | ok lets start | 21:23 |
yorkerlin | ok | 21:23 |
@HeikoS | Welcome to our GSoC 2014 celebration meeting. | 21:23 |
@HeikoS | We want to use this opportunity to thank everyone involved for their kick-ass work, we are extremely excited about the results this year and think that all project made significant contributions to Shogun. | 21:24 |
@HeikoS | The community of our little ML project is ever growing, and we really hope that a few of you guys will stay around afterwards -- you are MORE than welcome. | 21:24 |
@HeikoS | Any form of contribution is welcome: extending your/other projects, fixing bugs in your/other projects, writing more examples, blog posts, release help. | 21:24 |
@HeikoS | Agenda for today | 21:24 |
@HeikoS | -Project summaries (please every student/mentor pair generate some 3-5 sentences that summaries the project, including links to notebooks/examples/etc | 21:24 |
@HeikoS | -Blog posts about GSoC | 21:24 |
@HeikoS | -Release | 21:24 |
@HeikoS | -What next? | 21:24 |
@HeikoS | anyone has some other points? | 21:24 |
@HeikoS | anyone not asleep? :D | 21:25 |
@thoralf | Thanks you for the introduction. | 21:26 |
lambday | I can go first | 21:26 |
@HeikoS | ok then, I suggest we start with the project summaries | 21:26 |
@HeikoS | lambday: cool! | 21:26 |
lambday | (have some text ready) | 21:26 |
lambday | ok | 21:26 |
lambday | my project was about implementing a number of kernel hypothesis testing in shogun that would enable us to to perform a number of cutting edge kernel two-sample tests and kernel-independence tests to measure variable interactions in generic machine learning problems. | 21:26 |
@HeikoS | thoralf: do you wanna go next with abinash_panda, jiaolong ? | 21:27 |
lambday | The final target was to build a flexible framework using such measures for feature selection. | 21:27 |
lambday | In course, we faced some difficulties while implementing the algorithms and dealt with the issues theoretically as well as implementation-wise. | 21:27 |
lambday | What we have presently in shogun is a framework for kernel hypothesis testing flexible enough to incorporate new measures as well as the feature selection framework that can readily use those measures as a criterion for selecting best features among all. | 21:27 |
lambday | A few things are yet to be completed - which includes refactoring the hypothesis framework to incorporate streaming features support for kernel independence test (something that we discovered later that we will need). | 21:27 |
lambday | Another thing (wiking would like this) is that we'd want our feature selection algorithm to work with sparse features as well. | 21:27 |
@HeikoS | lambday: haha, indeed wiking would like this ! :) | 21:28 |
lambday | I am still working on the notebook for feature selection to be finished and hoping to push this by tomorrow. | 21:28 |
-!- hushell [~hushell@c-98-232-237-144.hsd1.or.comcast.net] has joined #shogun | 21:28 | |
lambday | HeikoS: yes :P | 21:28 |
@HeikoS | lambday: oh yeah that would be great | 21:28 |
@HeikoS | hushell: hi! | 21:28 |
lambday | In future I will continue with this work and also planning to join Heiko and Dino in London for further work (which includes leaving my job which I hate currently). | 21:28 |
hushell | HeikoS: hey!! | 21:28 |
* lambday done | 21:28 | |
@HeikoS | lambday: oh yeah we should talk about this a bit soon :) | 21:28 |
lambday | HeikoS: will do a PM | 21:28 |
@thoralf | Next one. Large-Scale Multilabel Project with abinash_panda. | 21:29 |
@HeikoS | thoralf: go for it! | 21:29 |
@thoralf | The project aimed on large-scale multilabel classification in SHOGUN. We succeeded in proving that we can learn the multilabel problem and implemented different approaches. | 21:30 |
@HeikoS | (the others can prepare text already) | 21:30 |
@thoralf | We actually didn't succedd in the "large-scale" part, but we already have some ideas ready how to improve perfomance of structured-output framework. | 21:30 |
-!- sonne|osx [~sonne@f052229200.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 21:30 | |
@HeikoS | thoralf: cool, details? | 21:30 |
@HeikoS | Saurabh7: can you be the next one? after thoralf is done? | 21:31 |
Saurabh7 | HeikoS, okie | 21:31 |
@thoralf | Overall complexity was quite high, so I'd say it's great for SHOGUN that we can advertize multilabel classification as a new feature. | 21:31 |
@thoralf | Details of how to improve SO framework would exceed this meeting. I guess I'm done here. ;) | 21:32 |
@HeikoS | thoralf: ok | 21:32 |
@HeikoS | Saurabh7: want to be next? | 21:32 |
Saurabh7 | yup | 21:32 |
@HeikoS | yorkerlin: can you also already prepare something? | 21:32 |
Saurabh7 | hi guys | 21:32 |
Saurabh7 | So as a part of my project I wrote detailed notebooks about various topics like SVMs, regression models (http://www.shogun-toolbox.org/static/notebook/current/Regression.html), introduction to ml with shogun (http://www.shogun-toolbox.org/static/notebook/current/Introduction.html), evaluation , etc. I have tried to organize them properly so as to benefit new users. | 21:32 |
Saurabh7 | Also I made various changes and improvements to the web demos adding features like model selection, usage of databases and more flexibility and coverage of shoguns capabilities e.g in GPs. | 21:32 |
@HeikoS | Saurabh7: the new web demos with GP are quite cool btw everyone, | 21:33 |
Saurabh7 | done :) | 21:33 |
@HeikoS | Saurabh7: wiking needs to help deploy this, but you can already clone the repo and play with it locally | 21:33 |
@HeikoS | Saurabh7: thanks! :) | 21:33 |
Saurabh7 | eah we should update those on the website i guess | 21:33 |
@HeikoS | Saurabh7: you should help us pushing this (otherwise it might take a while, its your work so you have a good motivation :) | 21:34 |
@HeikoS | yorkerlin: want to be next? | 21:34 |
yorkerlin | still writing on it | 21:34 |
-!- sonne|osx [~sonne@f052229200.adsl.alicedsl.de] has joined #shogun | 21:34 | |
@HeikoS | jiaolong: ? | 21:35 |
@thoralf | Approximate inference in SO framework (hushell and Jiaolong). ipython notebooks: | 21:35 |
@thoralf | https://github.com/shogun-toolbox/shogun/blob/develop/doc/ipython-notebooks/structure/Binary_Denoising.ipynb | 21:35 |
@thoralf | https://github.com/shogun-toolbox/shogun/blob/develop/doc/ipython-notebooks/structure/FGM.ipynb | 21:35 |
@thoralf | I can try a quick summary on this one as well. | 21:35 |
@HeikoS | thoralf: great ! | 21:36 |
@HeikoS | ok while people are writing, here is one for parijat and fernando | 21:36 |
@thoralf | I think this project was successful as well. The notebooks show different applications that Jiaolong made possible with our SO learning framework. | 21:36 |
hushell | Thanks thoralf :) | 21:36 |
jiaolong | Thanks | 21:36 |
@HeikoS | Parijat: "The summary of the work done in Fundamental ML algorithms in the second half of GSoC is as follows :1. Implement KD-Trees and Ball trees for fast NN search.2. Implement kernel density estimation using single tree as well as dual tree approach.3. Implement generic EM and generic mixture model classes. The CGaussian class needs to be refactored and other classes like that for multinomial distribution need to be implemented so that this whol | 21:37 |
@HeikoS | With orientations and seminars everyday, I am finding much less time to spend on Shogun these days. But as things settle down a bit, I should be able to contribute more often. " | 21:37 |
jiaolong | There is anoter one for scene segmentation: http://nbviewer.ipython.org/gist/Jiaolong/95633b7efd64bbd096f2 | 21:37 |
@HeikoS | everyone check out parijats stuff, very nice! | 21:37 |
@thoralf | jiaolong: Can you name the three (?) approximations you implemented? | 21:37 |
@HeikoS | jiaolong, thoralf, these notebooks should be merged asap btw | 21:37 |
jiaolong | HeikoS: yes, I am just coming back from two confernece, will finish them soon | 21:38 |
@HeikoS | jiaolong: great! | 21:38 |
hushell | We still have several PRs left, will be done shortly | 21:38 |
@HeikoS | hushell: great, yeah that should be done soon, before we forget | 21:38 |
@HeikoS | but lets talk about that later, | 21:39 |
jiaolong | HeikoS: this summer we just finish two inference algs: graph cuts, max-product LP relaxation | 21:39 |
jiaolong | HeikoS: yeah | 21:39 |
jiaolong | I would like to stay in shougun to improve current algorithms and also implements new algorithms. | 21:39 |
jiaolong | As I am also doing working on domain adaptation, I would like to contribute to domain adaptation projects in shogun as well. | 21:39 |
hushell | HeikoS: I'll keep them in mind :) | 21:39 |
@HeikoS | jiaolong: you are super welcome to do so. lambday just joined the core-dev team as you know :) | 21:39 |
@HeikoS | jiaolong: great!, in particular your research paper code might be useful | 21:40 |
@HeikoS | hushell: nice :) | 21:40 |
@HeikoS | ok cool all done? | 21:40 |
@HeikoS | yorkerlin? | 21:40 |
@HeikoS | kislay__: ? | 21:40 |
kislay__ | yeah ready. | 21:40 |
@HeikoS | kislay__: go! | 21:40 |
kislay__ | The project aimed at making a lucid interfacing between Shogun and OpenCV | 21:41 |
kislay__ | The first half consisted of integrating OpenCV's Mat datastructure with Shogun's SGMatrix. Did some changes in the FisherLDA class for implementing fisherfaces face rec. directly. | 21:41 |
kislay__ | The second half consisted mainly of showing cool computer vision applications using Shogun and OpenCV through notebooks. It comprised of Automatic Number Plate Recognition, Scene classification using SIFT, Active Appearance Model along with markdown documentations showcasing the C++ integration. This half was more fun as it consisted of mainly implementations. | 21:41 |
kislay__ | I realized that there's humongous possibility of doing things in Shogun related to ML and CV. I have a thesis this year and was really thinking of doing something that not only helps Shogun but also spruce up my thesis. I haven't decided the topic yet so will love to get your inputs for this. | 21:41 |
-!- sonne|osx [~sonne@f052229200.adsl.alicedsl.de] has quit [Quit: sonne|osx] | 21:42 | |
kislay__ | done i think. | 21:42 |
@HeikoS | kislay__: great! | 21:42 |
@HeikoS | kislay__: what about your open PRs? | 21:42 |
@HeikoS | kislay__: you sent an email on thise | 21:42 |
@HeikoS | we should try to get them merged soon, you should push everyone to do this :) | 21:42 |
kislay__ | yeah. I have to take care of some of them. | 21:42 |
@HeikoS | ok good! | 21:42 |
@HeikoS | so next one: yorkerlin | 21:43 |
yorkerlin | ok | 21:43 |
yorkerlin | The GSoC project I implemented several variational methods for GPC. The variational methods include: approximating the posterior covariance matrix exactly (for accuracy) assuming artificial structure of the posterior covariance matrix (for speed) However, there methods still NOT scalable to big data. I still working on refining the notebook but will attempt to finish it this week. | 21:43 |
yorkerlin | The next step is to implement large scale methods via stochastic variational approach and variable selection approach. (already read some papers and will implement them this year) The final goal is to extend to latent GP models and hierarchical latent models (eg, correlated topic models). | 21:43 |
@HeikoS | yorkerlin: yeah the notebook updates are important, please push them | 21:44 |
@HeikoS | yorkerlin: awesome, looking super forward to you to implement those things! | 21:44 |
@HeikoS | yorkerlin: we will talk to emt about all that | 21:44 |
yorkerlin | ok | 21:44 |
@HeikoS | I might submit a nips workshop on this | 21:44 |
@HeikoS | we will see | 21:44 |
@HeikoS | ok, did we forget any project? | 21:44 |
yorkerlin | cool! | 21:44 |
@HeikoS | lambday, lisitsyn ? | 21:45 |
khalednasr | HeikoS, I | 21:45 |
@HeikoS | khalednasr: ah yes | 21:45 |
@HeikoS | sorry | 21:45 |
@HeikoS | khalednasr: go for it :) | 21:45 |
khalednasr | I'll go :) | 21:45 |
khalednasr | The goal of the project was to implement the basic deep learning algorithms. The following was implemented: Standard neural networks, onvolutional nets, Restricted Boltzmann Machines, Deep Belief Networks, Deep autoencoders and Convolutional autoencoders. | 21:45 |
khalednasr | As documentation, I wrote notebooks and c++ examples that cover all the algorithms in the neuralnets module. | 21:45 |
khalednasr | I also did some work refactoring the linalg library and adding most of the methods needed by the neuralnets module to it. | 21:45 |
khalednasr | What remains to be done is modifying the neuralnets module to make it use the linalg library, so that it'll support running neural nets on the GPU. I'll try to find time to work on that sometime soon. | 21:45 |
@HeikoS | khalednasr: do the things already work again linalg? | 21:45 |
@HeikoS | khalednasr: that would be *amazing* | 21:46 |
@lisitsyn | yeah that's the crucial thing | 21:46 |
@HeikoS | khalednasr: in fact, if we get lambday's parallel/distributed framework running, we can parallelise things on GPU clusters | 21:46 |
@HeikoS | khalednasr: (you need some minor refactoring for that only) | 21:46 |
khalednasr | HeikoS, yeah that'd be good | 21:46 |
@HeikoS | khalednasr: but very exciting, I think we should aim to reproduce a large benchmark result, I can donate some time on a GPU cluster | 21:47 |
@HeikoS | khalednasr: so if you are interested, we can make Shogun be able to do that | 21:47 |
@HeikoS | khalednasr: but linalg is first | 21:47 |
khalednasr | yeah | 21:47 |
@HeikoS | khalednasr: can you also post your notebook links=? | 21:47 |
@HeikoS | ok, so I guess thats it, any comments? | 21:48 |
@HeikoS | if not, the second part of the meeting is quick: just me talking about a few things, please stop me and get a discussion going, if you feel like it | 21:48 |
@HeikoS | So second part of agenda: | 21:48 |
@HeikoS | -Blog posts about GSoC | 21:48 |
@HeikoS | We would like to encourage everyone of you to blog about GSoC. | 21:48 |
@HeikoS | This is a great way for us to get publicity. | 21:48 |
@HeikoS | It helps us doing the GSoC again next year: It helps us getting in, it makes more students apply, etc. | 21:48 |
@HeikoS | We would be EXTREMELY thankful if you could produce something *this* week. | 21:48 |
@HeikoS | (or next) | 21:49 |
@HeikoS | It does not have to be super polished - just write about your experience, and present some of your work. | 21:49 |
@HeikoS | social media would also be fine | 21:49 |
@HeikoS | even if just a few sentences | 21:49 |
@HeikoS | We would like to collect all posts and reference from a big one that we are writing. | 21:49 |
@HeikoS | The plan is to get it featured at http://google-opensource.blogspot.co.uk/ | 21:49 |
@HeikoS | See also the github issue https://github.com/shogun-toolbox/shogun/issues/2523 | 21:49 |
@HeikoS | comments? | 21:49 |
@HeikoS | lisitsyn, thoralf, sonney2k, wiking, abinash_panda, hushell, khalednasr, kislay__, lambday, pickle27, Saurabh7, yorkerlin ? | 21:50 |
yorkerlin | cool. will working on it next week since the notebook job. | 21:50 |
@HeikoS | yorkerlin: as said, it doesnt need to be super advanced | 21:50 |
@HeikoS | yorkerlin: just some content on Shogun and GSoC | 21:51 |
@HeikoS | Ok then next point! (wiking would be helpful here) | 21:51 |
@HeikoS | RELEASE. Shogun 3.4 | 21:51 |
@thoralf | HeikoS: Fine with writing a blog post here. | 21:51 |
@thoralf | HeikoS: One per project or a big one? | 21:51 |
@HeikoS | thoralf: cool, please share the link once done! | 21:51 |
@thoralf | LOL. | 21:51 |
@HeikoS | thoralf: both (coordinate the big one with me) | 21:51 |
@thoralf | HeikoS: Current status is "I like the idea" ;) | 21:51 |
@HeikoS | thoralf: Ill share a doc for the big one, but a small one on the projects would also be cool | 21:51 |
@HeikoS | haha | 21:52 |
@HeikoS | thoralf: greatly appreciated! | 21:52 |
@HeikoS | Release: | 21:52 |
@HeikoS | We aim for a release ASAP to put all your great work to the world (really important, and cool 8-) ). | 21:52 |
@HeikoS | For that, we need your help. | 21:52 |
@HeikoS | HELP | 21:52 |
@HeikoS | :) | 21:52 |
@HeikoS | First, make sure your project is in a release-able state (people will judge you on the quality of your work, so better clean again ;) ). | 21:52 |
hushell | also on personal webpages, facebook, twitter, google+ will be nice ;) | 21:52 |
@HeikoS | hushell: yes ! | 21:52 |
@HeikoS | Second, we got a list of things to solve for the release that are general minor (and major) issues all around Shogun. | 21:52 |
@HeikoS | Feel free to help us hacking on any of those. | 21:52 |
@HeikoS | Here is a list https://github.com/shogun-toolbox/shogun/milestones/Shogun%203.4 | 21:52 |
@HeikoS | Also this means: FEATURE FREEZE until we release. | 21:53 |
@HeikoS | Please do not send patches that extend Shogun before release, but rather fix some open problem. We count on you :) | 21:53 |
@HeikoS | lisitsyn, thoralf, sonney2k, wiking, abinash_panda, hushell, khalednasr, kislay__, lambday, pickle27, Saurabh7, yorkerlin we need all the hands we can for a release, any comments on that? | 21:53 |
@HeikoS | if not, lets go to the last point of the meeting | 21:54 |
@thoralf | HeikoS: Please let's start with a milestone on github. | 21:54 |
lambday | HeikoS: my first goal is to finish the notebook - then will write a NEWS about linalg which will be useful for the release | 21:54 |
hushell | what's the timeline for 3.4? | 21:54 |
@HeikoS | hushell: it would be good this month | 21:54 |
@HeikoS | thoralf: that link is a milestone | 21:54 |
lambday | HeikoS: also, was discussing something with lisitsyn about something - will discuss with that issue with you | 21:54 |
@HeikoS | thoralf: feel free to extend (there is an old clean-up release milestone) | 21:54 |
@thoralf | HeikoS: :) | 21:54 |
@HeikoS | thoralf: but only feasible stuff | 21:55 |
@HeikoS | we tend to put too much | 21:55 |
@HeikoS | this release is really GSoC feature, so just a few things that should work | 21:55 |
hushell | HeikoS: ok, the left behind PRs have to be hurried up | 21:55 |
@HeikoS | mainly binary packages, warnings, severe bugs, build, etc | 21:55 |
@HeikoS | hushell: yeah exactly | 21:55 |
@HeikoS | all PRs should be closed | 21:55 |
@HeikoS | hushell: I mean merged | 21:56 |
@HeikoS | lambday: ok | 21:56 |
@HeikoS | lambday: release related | 21:56 |
@HeikoS | ? | 21:56 |
@thoralf | HeikoS: Binary package for OSX shouldn't be part of the milestone. This issue may explode. ;) | 21:56 |
hushell | HeikoS: got it | 21:56 |
@HeikoS | thoralf: yep, I agree, feel free to remove it | 21:56 |
lambday | HeikoS: not exactly ... but future shogun goal related | 21:56 |
@HeikoS | lambday: ok | 21:56 |
@HeikoS | so really main focus is the release now | 21:57 |
@HeikoS | the world shall know about the great stuff we did | 21:57 |
@HeikoS | lisitsyn, thoralf, sonney2k, wiking, abinash_panda, hushell, khalednasr, kislay__, lambday, pickle27, Saurabh7, yorkerlin there is for example the task to remove all warnings from the buildbot | 21:57 |
@HeikoS | very easy, but lot of work | 21:57 |
@HeikoS | mostly documenting functions, fixing math, etc | 21:57 |
@HeikoS | static warnings another one | 21:57 |
@HeikoS | ok if no comments, then the last part of the agenda: | 21:58 |
@HeikoS | WHATS NEXT? (after release) | 21:58 |
@HeikoS | Stammtisch: We would love to get the Stammtisch going again. | 21:58 |
@HeikoS | This is every Monday evening, and everyone is welcome to talk about Shogun, and other things. | 21:58 |
@HeikoS | I will send emails about this. It would be great if you guys would hang out and if we get a discussion going on a regular basis. | 21:58 |
@HeikoS | comments on Stammtisch? | 21:58 |
@HeikoS | If not, the very last point | 21:59 |
@HeikoS | Finally, GSoC brought a few things to light, that would be great to continue working on. | 21:59 |
@HeikoS | In particular, the linalg framework would be a great great improvement if being used globally in all Shogun. | 21:59 |
@HeikoS | khalednasr, lambday, lisitsyn you guys shoud lead that btw | 21:59 |
@HeikoS | Everyone could help pushing this via adding new operations to the interface, and using them from algorithms. | 21:59 |
@HeikoS | I will soon add github issues on this. | 21:59 |
@lisitsyn | yeah I'll try to join that process again | 21:59 |
@HeikoS | lisitsyn: its not really about adding operations to the interface, and start porting our algorithms to that | 22:00 |
@HeikoS | Any other thoughts ideas? What do you think are things that would be good to work on this winter (before next GSoC) | 22:00 |
@lisitsyn | we need GWoC | 22:00 |
@lisitsyn | :D | 22:00 |
@thoralf | HeikoS: Regarding what's next. | 22:00 |
@thoralf | 1. Improving usability for "new" users. Providing installation scripts/binary packages for common platforms (Ubuntu/Debian, MacOSX, ?). | 22:00 |
@thoralf | Maybe this can be done after GSoC release within another release. I created a milestone 3.4.1 for this. This is open for discussion and changes. | 22:00 |
@thoralf | 2. Making SHOGUN more scalable (multilabel project showed that we still don't scale well). | 22:00 |
@thoralf | 2a. If we don't know how far we scale try "bigger" examples, not only small toy data sets. (Not neccessarily from tests) | 22:00 |
@thoralf | 2b. Investigate bordercases which don't scale and try to optimize SHOGUN on them. | 22:01 |
@HeikoS | thoralf: great points, maybe also add this to the shogun hack agenda | 22:01 |
@HeikoS | thoralf: great ideas, | 22:01 |
@thoralf | Link? | 22:01 |
@HeikoS | thoralf: I agree we really want to make shogun large scale | 22:01 |
@HeikoS | thoralf: we should have github issues for all this, and link them to the wiki | 22:01 |
@thoralf | HeikoS: Okay, regarding 1. we're on the same page. | 22:01 |
@HeikoS | thoralf: its here: https://github.com/shogun-toolbox/shogun/wiki | 22:01 |
@HeikoS | thoralf: https://github.com/shogun-toolbox/shogun/wiki/Roadmap-Shogun-2015-hack | 22:02 |
@thoralf | Regarding 2. we first should create awareness in the core team to encourage people to benchmark and optimize. | 22:02 |
@HeikoS | thoralf: very good point, | 22:02 |
@thoralf | (Not stopping after successful toy examples.) | 22:02 |
@HeikoS | gunnar and me talked a bit about that in fact | 22:02 |
@HeikoS | thoralf: yeah. totally | 22:02 |
@HeikoS | thoralf: and stop adding new methods, the red line for the next year should be improving existing things | 22:02 |
@HeikoS | i.e. faster/stable/usable | 22:03 |
@HeikoS | thoralf: would be great if you could document your thouhts, that makes it easier | 22:03 |
@HeikoS | thoralf: maybe you have an idea for a kind of development guideline that ensures scalability ? (like we did with the unit tests back then) ? | 22:03 |
@HeikoS | lambday: you also had a point on whats next? | 22:04 |
@HeikoS | I guess we can close the meeting and then continue discussing | 22:04 |
@HeikoS | If there are no other points, we would like to heartly thank you all again for your great work. We love you all and hopefully see you around :) | 22:04 |
@HeikoS | Meeting is closed. | 22:04 |
lambday | HeikoS: something that I thought of while trying to refactoring hypothesis framework | 22:04 |
lambday | HeikoS: in many many places in shogun, we need to rely on some combinatorial combinations of componets | 22:05 |
lambday | for which we presently mostly rely on virtual calls | 22:05 |
lambday | I was thinking if we can do better than that | 22:05 |
@thoralf | HeikoS: It's hard to automate benchmarking. I think people should try to apply their stuff to bigger data and then use profiling to find "hotspots". | 22:06 |
lambday | the alternative being relying on highly templated design | 22:06 |
@thoralf | Then start optimizing the hotspots. | 22:06 |
lambday | the only issue with this is that we might have difficulties while exposing to modular interaces | 22:06 |
@HeikoS | thoralf: yeah, but I mean something like, you always have to include a benchmark or so, because people quickly forget those things | 22:06 |
lambday | for that I was thinking that maybe we can use CSGObjects as wrappers mostly | 22:06 |
@HeikoS | lambday: cool, so whats the gain there? speed? | 22:06 |
lambday | and send the most time consuming parts to c++ internals | 22:07 |
lambday | HeikoS: exactly | 22:07 |
@HeikoS | lambday: in fact, that could be great as a motivation to clean shoguns internals up | 22:07 |
lambday | HeikoS: just have a look at this simple example - https://gist.github.com/lambday/73b9e68ec08e2da25967 | 22:07 |
@HeikoS | lambday: would you mind documenting this a bit? like basic idea, example, what has to be done, etc? | 22:07 |
lambday | HeikoS: yeah | 22:07 |
@HeikoS | lambday: and put it in the wiki, I mean, write a gitrhub issue and link it in the wiki for the hack | 22:07 |
lambday | HeikoS: I was already planning to write something on this on the refactoring hypothesis frameworks issue | 22:08 |
lambday | HeikoS: for example, it sucks that we can't use sparse features with gaussian kernels currently | 22:08 |
lambday | because using existing stuffs its tough to reimplement that | 22:08 |
-!- sonne|osx [~sonne@f052229200.adsl.alicedsl.de] has joined #shogun | 22:08 | |
lambday | I think we can do better | 22:08 |
@HeikoS | lambday: so thats one order of magnitude? | 22:09 |
@HeikoS | lambday: yeah | 22:09 |
@HeikoS | lambday: wow, that actually would mean we could fe-factor a lot of the internals right? | 22:09 |
lambday | HeikoS: also, plugin framework that lisitsyn proposed is essential I think | 22:09 |
@HeikoS | lambday: these things are the ones we should do within the next year, really get Shoguns c++ code modernized, and faster this way | 22:09 |
lambday | HeikoS: yeah just an idea though | 22:09 |
@lisitsyn | guys as for plugins framework | 22:09 |
@HeikoS | lambday: also we can think about many existing problems | 22:09 |
@HeikoS | as serialisation etc | 22:10 |
lambday | HeikoS: yeah | 22:10 |
@lisitsyn | I need some help with it I think | 22:10 |
@HeikoS | lambday: can you write a little motivation for all this? | 22:10 |
@HeikoS | so that everyone is aware? | 22:10 |
@HeikoS | lambday: this is also super useful for grant applications (like for the hack | 22:10 |
@HeikoS | lisitsyn: same there, | 22:10 |
lambday | so maybe we can spend until next GSoC more on these engineering issues than ML stuffs | 22:10 |
@HeikoS | lambday: totally | 22:10 |
@lisitsyn | HeikoS: same with what? :) | 22:10 |
@HeikoS | lisitsyn: I am willing to help | 22:10 |
@lisitsyn | ok just a few first steps then | 22:11 |
@HeikoS | lisitsyn: would be good to write down your ideas in a compressed way, and then link them in the wiki I put up | 22:11 |
lambday | HeikoS: lisitsyn will have to help in this :D | 22:11 |
@HeikoS | lisitsyn: very useful for getting money for the hack (if we have discussions and concrete ideas) | 22:11 |
@lisitsyn | lets we three HeikoS lambday agree on naming | 22:11 |
@lisitsyn | then I put it into shougn | 22:11 |
@lisitsyn | as shogun org | 22:11 |
@lisitsyn | at github | 22:11 |
@lisitsyn | we will have some kind of separate project :D | 22:11 |
@lisitsyn | so if anyone wants some kind of plugin based swig library - just use our core thing | 22:12 |
@HeikoS | lisitsyn: everyone wants this | 22:12 |
@HeikoS | lisitsyn: you know, at the meeting in NY, everyone was jealous that we can do the interface thing | 22:12 |
@HeikoS | lisitsyn: would be good to have a separate project, I already talked to s?ren about this last year | 22:13 |
@HeikoS | lisitsyn: please write down | 22:13 |
@HeikoS | lisitsyn: all good for the hack | 22:13 |
@HeikoS | lisitsyn: we can do all this in a week | 22:13 |
@lisitsyn | HeikoS: no I just propose to move it into github org already | 22:13 |
-!- Mathnerd314 [~quassel@dsl-206-53-27-58.den.pcisys.net] has quit [Read error: Connection reset by peer] | 22:13 | |
@lisitsyn | wiki is ok as well | 22:13 |
@lisitsyn | but I want to push it a bit other way as well | 22:13 |
@HeikoS | lisitsyn: just for internal synchronisation etc | 22:13 |
@lisitsyn | I need a name :D | 22:14 |
@lisitsyn | HeikoS: lambday: | 22:14 |
@HeikoS | lisitsyn: I will think of something | 22:14 |
@lisitsyn | currently its aer :D | 22:14 |
@HeikoS | lisitsyn: applications of this are huge, stan people for example want this too | 22:14 |
lambday | we gotta be the best in the field, men!! | 22:14 |
@HeikoS | lisitsyn: john langford told me: "Basically you have everything that every other project would like to have" | 22:14 |
lambday | lisitsyn: name... bad person to ask | 22:14 |
@HeikoS | lambday: yes sir! | 22:14 |
-!- Saurabh7 [~guestie_@103.225.100.51] has quit [Quit: Leaving] | 22:15 | |
@lisitsyn | okay then it stays aer yet :D | 22:15 |
lambday | lisitsyn: aer++ | 22:15 |
-!- Mathnerd314 [~quassel@dsl-206-53-27-58.den.pcisys.net] has joined #shogun | 22:15 | |
@HeikoS | lisitsyn: please write a bit about the general idea in the wiki, cannot say that often enough :) | 22:15 |
@HeikoS | aer ++ haha | 22:15 |
@HeikoS | c++^2 | 22:16 |
lambday | latin for air, eh? :P | 22:16 |
-!- khalednasr [~k.nasr92@41.69.136.192] has quit [Quit: Leaving] | 22:16 | |
@HeikoS | lambday: I will check your c++ listings | 22:16 |
lambday | lisitsyn: so we'll start a whole new branch of shogun, refactoring bits and pieces as we go along | 22:16 |
@thoralf | HeikoS: I put my ideas to the hack page. | 22:16 |
@lisitsyn | https://github.com/shogun-toolbox/aer | 22:17 |
@lisitsyn | ok guys | 22:17 |
@lisitsyn | ohh it is outdated :D | 22:17 |
lambday | HeikoS: just a small example that is - to demonstrate that we can do better than java style interfaces using c++ templates :D | 22:17 |
-!- kislay__ [~Abhijeet@117.239.94.112] has quit [Quit: Leaving] | 22:17 | |
@lisitsyn | ok guys be back later | 22:19 |
@lisitsyn | I'll try to update you on this thing | 22:19 |
lambday | I am pretty psyched to get these stuffs working before the hackathon | 22:19 |
@HeikoS | lisitsyn: cool! | 22:19 |
@HeikoS | see you :) | 22:19 |
@HeikoS | lambday: awesome :) | 22:19 |
@HeikoS | yeah me too | 22:19 |
lambday | and there we can integrate the whole thing and have a super cool release | 22:19 |
@HeikoS | I am psyched about the hackathon | 22:19 |
@HeikoS | just need to get the money | 22:20 |
@HeikoS | and for that we need documented discussions | 22:20 |
@HeikoS | lambday: yeah agreed | 22:20 |
lambday | absolutely | 22:20 |
@HeikoS | lambday: ok cool | 22:20 |
@HeikoS | lambday: gotta go now, lets keep the stammtisch going, we should discuss those stuff at least once a week | 22:21 |
lambday | I'm gonna start this discussion by putting a few of the ideas for hypothesis testing framework | 22:21 |
lambday | HeikoS: absolutely | 22:21 |
@HeikoS | lambday: sweet | 22:22 |
@HeikoS | lambday: I will be around a bit more from now on | 22:22 |
lambday | if we have a working demo inside shogun that we can do this - then that's a good point to start future discussions | 22:22 |
@HeikoS | I agree | 22:22 |
lambday | HeikoS: me too | 22:22 |
@HeikoS | lambday: like the linalg stuff | 22:22 |
lambday | HeikoS: yeah | 22:22 |
lambday | thanks to khaled for the brilliant stuffs he added | 22:23 |
@HeikoS | lambday: +1 | 22:23 |
@HeikoS | khaled++ | 22:23 |
@HeikoS | lambday: I gotta go now (dinner) | 22:23 |
@HeikoS | lambday: lets chat this week! | 22:23 |
@HeikoS | I will be online! | 22:23 |
lambday | HeikoS: alright | 22:23 |
lambday | HeikoS: I'll try to chalk out a few things by then | 22:24 |
@HeikoS | lambday: cool | 22:24 |
@HeikoS | lambday: lets talk at stammtisch on monday eve latest | 22:24 |
lambday | HeikoS: alright. | 22:25 |
@thoralf | Bye! | 22:27 |
-!- thoralf [~thoralf@91.65.138.205] has quit [Quit: Konversation terminated!] | 22:27 | |
-!- PirosB3 [~pirosb3@host132-36-dynamic.9-87-r.retail.telecomitalia.it] has quit [Quit: PirosB3] | 22:30 | |
-!- pickle27 [~AndChat79@24.114.82.188] has quit [Quit: Bye] | 22:40 | |
lambday | lisitsyn: please write something about the ideas on the plugin framework :P | 22:47 |
lambday | lisitsyn: wiking suggested that we can take inspiration from other existing libraries as well | 22:48 |
-!- HeikoS [~heiko@05439144.skybroadband.com] has quit [Quit: Leaving.] | 23:09 | |
-!- yorkerlin [18d4da3a@gateway/web/freenode/ip.24.212.218.58] has quit [Ping timeout: 246 seconds] | 23:38 | |
-!- pickle27 [~pickle27@192-0-136-118.cpe.teksavvy.com] has joined #shogun | 23:41 | |
-!- PirosB3 [~pirosb3@host132-36-dynamic.9-87-r.retail.telecomitalia.it] has joined #shogun | 23:52 | |
-!- HeikoS [~heiko@05439144.skybroadband.com] has joined #shogun | 23:53 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 23:53 | |
--- Log closed Sun Sep 07 00:00:31 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!