--- Log opened Sat Mar 18 00:00:09 2017 | ||
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Remote host closed the connection] | 00:00 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 00:21 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 256 seconds] | 00:25 | |
-!- Ting_ [52053693@gateway/web/freenode/ip.82.5.54.147] has left #shogun [] | 00:59 | |
-!- elenanis [4f67600c@gateway/web/freenode/ip.79.103.96.12] has joined #shogun | 01:36 | |
-!- elenanis [4f67600c@gateway/web/freenode/ip.79.103.96.12] has quit [Quit: Page closed] | 02:12 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 02:52 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 260 seconds] | 02:57 | |
-!- neha [ca8e6c9b@gateway/web/freenode/ip.202.142.108.155] has joined #shogun | 03:14 | |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-cpxnefbzuqtqotki] has joined #shogun | 03:27 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 05:00 | |
@sukey | Pull Request #3710 "Fixed broken links in ipython notebooks" synchronized by deveshnag1 - https://github.com/shogun-toolbox/shogun/pull/3710 | 05:01 |
---|---|---|
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 258 seconds] | 05:05 | |
-!- neha [ca8e6c9b@gateway/web/freenode/ip.202.142.108.155] has quit [Ping timeout: 260 seconds] | 05:32 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 06:21 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 246 seconds] | 06:25 | |
-!- goksinen_ [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 07:02 | |
-!- goksinen_ [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 264 seconds] | 07:07 | |
-!- nagdev [dce39546@gateway/web/freenode/ip.220.227.149.70] has joined #shogun | 08:49 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 09:03 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 260 seconds] | 09:08 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 09:43 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 260 seconds] | 09:48 | |
mikeling | wiking: hi, do you got time to review https://github.com/shogun-toolbox/shogun/pull/3711/files? Does it looks better to you? | 10:21 |
mikeling | I will address it if doesn't | 10:21 |
mikeling | :) | 10:21 |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun | 10:24 | |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has left #shogun [] | 10:24 | |
@wiking | mikeling, ah sorry man | 10:26 |
@wiking | i was bbusy lately | 10:26 |
@wiking | but i'll take a look in the next 24 hours | 10:26 |
@wiking | its a good idea | 10:26 |
@wiking | but i think we could leverage more | 10:27 |
@wiking | gtest's features | 10:27 |
@wiking | (see fixtures) | 10:27 |
mikeling | wiking: yeah, I know gtest is a great feature, but I haven't read it before. Let's me check its document then :) | 10:28 |
-!- Ting_ [52053693@gateway/web/freenode/ip.82.5.54.147] has joined #shogun | 11:03 | |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun | 11:18 | |
@sukey | Pull Request #3722 "Lock free InputParser and ParseBuffer" opened by tdjogi010 - https://github.com/shogun-toolbox/shogun/pull/3722 | 11:26 |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 11:44 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 246 seconds] | 11:49 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has joined #shogun | 12:01 | |
-!- Ting_ [52053693@gateway/web/freenode/ip.82.5.54.147] has quit [Quit: Page closed] | 12:05 | |
-!- goksinen [~goksinen@rrcs-50-75-193-138.nyc.biz.rr.com] has quit [Ping timeout: 256 seconds] | 12:06 | |
-!- Chris_ [8602a488@gateway/web/freenode/ip.134.2.164.136] has joined #shogun | 12:10 | |
-!- Chris_ [8602a488@gateway/web/freenode/ip.134.2.164.136] has quit [Ping timeout: 260 seconds] | 13:06 | |
-!- Chris___ [bc6266bf@gateway/web/freenode/ip.188.98.102.191] has joined #shogun | 13:29 | |
Chris___ | Hi, I would like to do MKL (2 Gaussian Kernels and 1 LocalAlignmentStringKernel, SVRlight). However during grid search after some iterations I get the following error when calling train() on the MKLClassification object: SystemError: [ERROR] In file /Applications/shogun/src/shogun/classifier/mkl/MKL.cpp line 848: Assertion R >= 0 failed! | 13:33 |
Chris___ | I found some related issues in the IRC logs but the fixes suggested there didn't help. | 13:34 |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has quit [Ping timeout: 260 seconds] | 14:27 | |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has joined #shogun | 14:41 | |
-!- geektoni [~geektoni@93-34-234-212.ip52.fastwebnet.it] has left #shogun [] | 14:41 | |
CaBa | heya | 15:04 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 16:00 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:00 | |
-!- Chris___ [bc6266bf@gateway/web/freenode/ip.188.98.102.191] has quit [Ping timeout: 260 seconds] | 16:08 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 260 seconds] | 16:18 | |
@sukey | Pull Request #3719 "Implement QR solver in linalg library" synchronized by micmn - https://github.com/shogun-toolbox/shogun/pull/3719 | 16:35 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 16:37 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 16:37 | |
-!- stargazer [~Gautam@116.75.186.133] has joined #shogun | 16:42 | |
stargazer | hey folks! | 16:50 |
stargazer | I want to work with shogun in this GSoC | 16:50 |
stargazer | i'd like to discuss my idea with you guys | 16:51 |
stargazer | any of the mentors available now? | 16:51 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 246 seconds] | 16:52 | |
-!- abhinav [~abhinav@14.192.52.215] has joined #shogun | 17:07 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 17:18 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 17:18 | |
stargazer | HeikoS: you're gonna be mentoring the application of Shogun project, right? | 17:19 |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 260 seconds] | 17:46 | |
-!- abhinav [~abhinav@14.192.52.215] has quit [Quit: Leaving] | 18:15 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has joined #shogun | 18:25 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 18:25 | |
@HeikoS | stargazer: yes I will be mentoring | 18:31 |
stargazer | oh great | 18:49 |
stargazer | HeikoS can we discuss my idea here? | 18:50 |
@HeikoS | stargazer: sure! | 18:50 |
stargazer | HeikoS: | 18:50 |
stargazer | have you heard about data.gov.in ? | 18:51 |
stargazer | it is a website that has a lot of datasets open to public access by indian government | 18:51 |
stargazer | it has 73,067 RESOURCES | 18:51 |
@HeikoS | I dont know this one | 18:51 |
stargazer | 1,039 VISUALIZATIONS and 444 APIs | 18:52 |
@HeikoS | but I know of similar sites in the uk | 18:52 |
stargazer | it is a very new initiative by our government and has a lot of untapped potential | 18:52 |
@HeikoS | yes I guess so | 18:53 |
stargazer | i think we just have to mix and match with any datasets/domains and it can be a very useful and cool application of ML toolkit like shogun | 18:53 |
@HeikoS | yes definitely | 18:53 |
@HeikoS | this is exactly the kind of stuff we want to do in the applied projects | 18:53 |
@HeikoS | you saw the wiki I guess? | 18:54 |
stargazer | can you guide me so as to how i should proceed? | 18:54 |
stargazer | yeah | 18:54 |
@HeikoS | well, read the wiki :D | 18:54 |
stargazer | as soon as i saw it, this popped up in my mind | 18:54 |
stargazer | actually i read it | 18:54 |
stargazer | heres a doubt | 18:54 |
@HeikoS | so these applied projects require a lot of selt initiative | 18:54 |
@HeikoS | so the best way to proceed is to show that | 18:54 |
@HeikoS | and also you need a really clear vision of what you want to do | 18:55 |
@HeikoS | just "applying ML to dataset" is not enough | 18:55 |
stargazer | I'm gonna be 'using' the toolkit for most of the part, instead of working on the source of the toolkit itself | 18:55 |
stargazer | so is it necessary still, to contribute before hand to the repo? | 18:55 |
stargazer | yeah.. i'll need some time to figure out what it is that i would precisely want to achieve | 18:56 |
stargazer | can you tell me any 'for instance'? | 18:56 |
@HeikoS | no for the applied project there is no need to contribute something | 18:57 |
@HeikoS | but you definitely have to show something concrete | 18:57 |
@HeikoS | like a proof of concept of what you want to do | 18:57 |
stargazer | okay | 18:57 |
CaBa | re | 18:57 |
@HeikoS | the question is: what do you want to achieve with a project? | 18:57 |
@HeikoS | why would we pay you for 3 months to do a project, you need to convince us, and this works best with concrete things that you show | 18:58 |
@HeikoS | doesnt need to be code, but something in the spirit of what you want to achieve | 18:58 |
@HeikoS | it is some work to spell this out | 18:58 |
@HeikoS | but can be really fun | 18:58 |
stargazer | one potential idea, is to analyse the efficiency of our education system | 18:58 |
@HeikoS | but as we said, we cannot tell you what to do here, it needs to come from you | 18:59 |
stargazer | yeah | 18:59 |
@HeikoS | stargazer: that sounds like a great idea | 18:59 |
@HeikoS | but now you need to think: how would you actually do that? | 18:59 |
@HeikoS | what would be the deliverable of such a project? | 18:59 |
stargazer | yes | 18:59 |
stargazer | in ML projects, what deliverables are generated usually? | 18:59 |
stargazer | i have never seen one in action before | 19:00 |
stargazer | is it visualisation of datasets and predictions? | 19:00 |
@HeikoS | well, think of one :) | 19:01 |
@HeikoS | yeah visuals are cool | 19:01 |
@HeikoS | demos | 19:01 |
@HeikoS | blog posts | 19:01 |
@HeikoS | insights | 19:01 |
@HeikoS | can be anything really | 19:01 |
-!- witness_ [uid10044@gateway/web/irccloud.com/x-pskggscgommuwidu] has joined #shogun | 19:01 | |
@HeikoS | but has to be somehow useful, or has to be interesting to others | 19:01 |
@HeikoS | but again | 19:01 |
@HeikoS | that depends on what you want to do | 19:01 |
stargazer | blog posts is a requirement if i'm not mistaken. so they'll definitely be there | 19:02 |
@HeikoS | once that is clear, it is easy to define deliverables | 19:02 |
stargazer | okay | 19:02 |
stargazer | great | 19:02 |
@HeikoS | if you make something interactive, a docker image for reproducing that is necessary | 19:03 |
stargazer | another idea, could be to predict economic growth of various states | 19:03 |
@HeikoS | yeah good one | 19:04 |
stargazer | yeah, sure. | 19:04 |
@HeikoS | great ideas | 19:04 |
@HeikoS | just make sure to chalk it out nicely | 19:04 |
@HeikoS | and have a clear concept of what you want to do | 19:04 |
stargazer | let me make sure that i get this clear then, | 19:04 |
stargazer | i will be building and training an ML model made with shogun, with data available from the given source, and producing the output of that ML model in the enf | 19:05 |
stargazer | end | 19:05 |
@HeikoS | all this will be part of a project | 19:05 |
@HeikoS | BUT | 19:05 |
@HeikoS | on its own, that is not nearly enough to justify a project | 19:06 |
@HeikoS | takes 1 day | 19:06 |
@HeikoS | GSoC is 90days | 19:06 |
stargazer | yeah | 19:06 |
stargazer | :') | 19:06 |
stargazer | to add interactivity, we can make a game out of it maybe | 19:06 |
@HeikoS | maybe | 19:07 |
stargazer | like the project will be able to predict the outcomes of any conditions given by users, such as rainfall, famine, etc | 19:07 |
stargazer | on a map like gui | 19:07 |
stargazer | there can be many qualifiers like that | 19:07 |
@HeikoS | not sure what you have in mind there exactly, but interactive things on state data is extremely cool | 19:07 |
stargazer | lets say that after reading past data, our project predicts 3.5% increase in growth economically | 19:08 |
stargazer | now user says but what if the rainfall is extremely high this year, or for coming 3 years, how can that affect the economy? | 19:09 |
stargazer | something on these lines | 19:09 |
@HeikoS | maybe yes, but much more structured | 19:11 |
@HeikoS | but I see what you mean | 19:12 |
@HeikoS | could work | 19:12 |
stargazer | yeah | 19:12 |
CaBa | HeikoS: are you familiar with the MKL code? does this sound familiar: https://gitlab.unique-internet.de/snippets/16 | 19:12 |
stargazer | is it okay to allocate time for dataset research in my GSoC time? | 19:12 |
@HeikoS | stargazer: nope I dont think that is a good idea | 19:14 |
@HeikoS | CaBa: a bit | 19:14 |
@HeikoS | CaBa: never seen that | 19:14 |
@HeikoS | MKL is not maintained anymore sadly :( | 19:14 |
stargazer | okay | 19:15 |
@HeikoS | the guy who wrote it all is not active anymore | 19:15 |
CaBa | HeikoS: mostly sörens code i assume? | 19:15 |
@HeikoS | stargazer: generally, GSoC is the time to DO, befre that is the time to PLAN and research | 19:15 |
@HeikoS | CaBa: yeah | 19:15 |
stargazer | true | 19:17 |
stargazer | but for something like ML and AI based, datasets are very vital and if they're chosen wrongly, the whole project wouldn't make sense | 19:18 |
stargazer | also, can we use a combination of shogun and Java? | 19:18 |
CaBa | HeikoS: if you have time, please take a look whether you can give #3721 a go. I kept the old PR for reference of the problem, but I think the new one is the safer solution and it requires no further fixing... | 19:19 |
@HeikoS | CaBa: yeah I want to check that thing soon | 19:19 |
CaBa | cool | 19:19 |
@HeikoS | stargazer: shogun and java is fine and actually would be cool to use something non python | 19:19 |
stargazer | that's great | 19:19 |
stargazer | because i have a solid java background | 19:20 |
@HeikoS | stargazer: I agree datasets are vital, and also extremely vital for a GSoC project, that is why this needs to be sorted our in your application | 19:20 |
@HeikoS | stargazer: usually, one picks the language that suits the purpose most | 19:20 |
CaBa | HeikoS: I chose a silly new filename for the unit test btw, maybe this can better be added to the SGObject test file. that already contains some CDynamicArray tests... anyways, i'm sure you'll complain when you check the PR ;) | 19:20 |
stargazer | agreed | 19:20 |
@HeikoS | rather than finding a prupose for ones language | 19:20 |
@HeikoS | CaBa: yeah for sure Ill moan :D | 19:20 |
stargazer | i have not done ML before so all i've heard is Python is the way for such projects | 19:21 |
stargazer | i have worked with python a bit but nothing outstanding | 19:22 |
stargazer | HeikoS i have to sleep rn unfortunately, but we'll definitely talk later | 19:22 |
stargazer | you have been a great help | 19:22 |
@HeikoS | stargazer: cool, talk to you later then! | 19:22 |
-!- stargazer [~Gautam@116.75.186.133] has quit [Quit: Leaving] | 19:22 | |
-!- mikeling [uid89706@gateway/web/irccloud.com/x-cpxnefbzuqtqotki] has quit [Quit: Connection closed for inactivity] | 19:30 | |
-!- HeikoS [~heiko@host-92-0-178-129.as43234.net] has quit [Ping timeout: 260 seconds] | 20:29 | |
-!- witness_ [uid10044@gateway/web/irccloud.com/x-pskggscgommuwidu] has quit [Quit: Connection closed for inactivity] | 22:18 | |
CaBa | wiking: ping | 23:14 |
-!- zxtx [~zxtx@sunnythinking.org] has quit [Quit: ZNC - http://znc.in] | 23:36 | |
-!- neha [ca8e6ccb@gateway/web/freenode/ip.202.142.108.203] has joined #shogun | 23:38 | |
--- Log closed Sun Mar 19 00:00:10 2017 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!