13:01:09 #startmeeting NeuroFedora - 2020-11-16 13:01:09 Meeting started Mon Nov 16 13:01:09 2020 UTC. 13:01:09 This meeting is logged and archived in a public location. 13:01:09 The chair is bt0. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:01:09 Useful Commands: #action #agreed #halp #info #idea #link #topic. 13:01:09 The meeting name has been set to 'neurofedora_-_2020-11-16' 13:01:19 #meetingname neurofedora 13:01:19 The meeting name has been set to 'neurofedora' 13:01:54 #chair FranciscoD MeWjOr iztok85 13:01:54 Current chairs: FranciscoD MeWjOr bt0 iztok85 13:02:14 #topic Today's Agenda 13:02:33 #info introductions and roll call 13:02:33 #info Tasks from last bi-week's meeting: 13:02:33 #info Open Pagure tickets 13:02:33 #info Open NeuroFedora bugs 13:02:33 #info Koschei packages check 13:02:33 #info CompNeuro lab compose status check for F33/F34 13:02:35 #info Neuroscience query of the week 13:02:37 #info Next meeting day, and chair 13:02:39 #info open floor 13:02:50 #topic introductions and roll call 13:03:44 #info FAS: ankursinha, Ankur Sinha, IRC: FranciscoD ; NeuroFedora, packaging, Fedora-join, Ask Fedora etc. UTC+1 (Europe/London) 13:04:33 #info Alberto Rodriguez Sanchez, FAS: bt0dotninja, IRC: bt0; CommOps, NeuroFedora, Marketing, fedora-join, mindshare, etc. ; UTC -6 13:05:40 #info Aniket Pradhan, FAS: major, IRC/Telegram: MeWjOr NeuroFedora, Packaging, random stuff; UTC+5:30 13:06:18 I will wait a couple of minutes more until the next topic :) 13:06:54 #chair AarKiMos 13:06:54 Current chairs: AarKiMos FranciscoD MeWjOr bt0 iztok85 13:07:23 Hey there AarKiMos 13:07:55 hello, 13:08:17 Welcome AarKiMos 13:08:18 I am a newcomer, just observing for now 13:08:28 :D 13:08:36 AarKiMos++ 13:08:54 aarkimos++ 13:08:57 FranciscoD: Karma for aarkimos changed to 1 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:09:01 there we go :) 13:09:12 Ok, Next topic (aarkimos++ too) 13:09:30 aarkimos++ 13:09:30 bt0: Karma for aarkimos changed to 2 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:09:31 aarkimos++ 13:09:33 MeWjOr: Karma for aarkimos changed to 3 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:09:37 thanks 13:09:41 #topic Task from the last meeting 13:09:50 #link https://meetbot.fedoraproject.org/teams/neurofedora/neurofedora.2020-11-02-13.01.html 13:11:03 #info MeWjOr Work over the update for python-mne 13:11:25 Ahh... it's still a WIP 13:11:38 cool, next one 13:11:42 they just released a 0.21.2, will try if that fixes stuff 13:11:45 #info MeWjOr Fix the deps for python-odml 13:12:02 stuck with license on that :P 13:12:07 :P 13:12:24 bt0: best to keep re-doing the action command too, so we have them again in the log next time? 13:12:32 sure 13:13:10 #action the awesome MeWjOr Work over the update for python-mne 13:13:30 #action MeWjOr Fix the deps for python-odml 13:13:37 * MeWjOr blushes 13:13:43 #info FranciscoD Checkout nest at the next meeting if its building fine or not 13:14:18 it's building now, was something to do with koji 13:14:39 ok, re action it 13:14:50 no, solved :) 13:14:55 awesome 13:15:04 sorry, I meant "it builds now" 13:15:06 :) 13:15:23 #info alciregi check on python-pydicom package 13:16:11 Alessio ate you here? 13:16:25 (edited) Alessio, are you here? 13:17:19 maybe no, re action 13:17:35 +1 13:17:36 #action alciregi check on python-pydicom package 13:18:05 the next one is good 13:18:10 #info everyone Think over some ideas as to what we can present at FOSDEM 21 13:18:33 I was thinking perhaps we could show how to use fedora containers, and install stuff on them to work 13:18:49 🤔 13:19:01 as another use-case for NeuroFedora 13:19:08 Interesting thought, FranciscoD 13:19:13 sounds cool 13:19:33 I was not able to get the time to think much about it 13:19:36 that'll help with reproducibility---people can make fedora based docker containers etc 13:19:52 MeWjOr: i was looking at updating the docker page on the docs and thought this was newish 13:20:31 I can make a parallel process with podman 13:20:33 umm, yes... we do have a tracker to make a docker image for NeuroFedora 13:20:50 yeh, we don't necessarily need a docker image yet 13:21:00 just pull fedora:33, and then install stuff on it etc? 13:21:02 less work 13:21:04 :D 13:21:06 actually we can use the same Docker File to generate both images 13:21:25 docker and podman 13:21:28 +1 13:21:51 and it is possible to have container images for packages, but that's lots of work---needs a review etc. 13:22:05 I don't know much about docker images, but both sound plausible 13:22:15 sure 13:22:48 MeWjOr: you just "pull" the container, and then you can run it like a vm and do whatever you want pretty much 13:23:04 and you can make customised "images" etc 13:23:24 I can do a quick demo next meeting? 13:23:28 we can propose a base docker-file and fork it to generate the new images 13:23:43 +1 13:23:55 let me do it 13:24:03 it can have the same packages as the comp-neuro installer image perhaps 13:24:03 +1 works great 13:24:07 minus the workstation etc? 13:24:35 #action bt0 will generate a base docker-file for the neurofedora Docker/podman image 13:25:17 sounds good? 13:25:21 +1 13:25:45 ++ 13:26:48 awesome I will work on it and update the ticket #374 13:27:03 #link https://pagure.io/neuro-sig/NeuroFedora/issue/374 13:27:32 another ideas for the FOSDEM 21? 13:28:15 not too many---may be a dev session on our pipeline etc 13:29:05 #idea Dev session on FOSDEM 21 13:29:32 Ok, next topic 13:29:36 #topic Open Pagure tickets 13:29:43 #link https://pagure.io/neuro-sig/NeuroFedora/issues?status=Open&tags=S%3A+Next+meeting 13:29:53 #info Ticket 318 NeuroFedora reading/listening list 13:30:01 #link https://pagure.io/neuro-sig/NeuroFedora/issue/318 13:30:08 We have only this one 13:31:50 but noting to do (ok reading of course, but noting actionable) 13:32:25 next topic :) 13:32:28 #topic Open NeuroFedora bugs 13:32:37 #link https://tinyurl.com/neurofedora-bugs 13:33:07 I still see the mne bug there :P 13:33:08 many new bugs 13:33:57 thanks MeWj0r++ 13:34:09 MeWjOr++ 13:34:09 bt0: Karma for major changed to 2 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:34:16 :) 13:34:16 :D 13:35:54 I don't see urgent tickets here. do you? 13:36:08 some simple updates imo 13:36:15 Jorge has left the Telegram Group. 13:36:15 and FTBFS >.< 13:36:59 :( 13:37:42 next topic... 13:38:02 #topic Koschei packages check 13:38:12 #link https://koschei.fedoraproject.org/groups/neuro-sig 13:39:51 I see neurord, python-lfpy, python-missingno, pydicom 13:40:18 how can i help here? 13:41:50 looking logs or something? 13:42:24 wait... missingno was a dep for chaospy... 13:42:40 and it has a new update probably 13:42:54 it might turn out to be a mess in the future 13:43:06 lol, yes 13:43:36 I'll probably look into it if time allows 13:44:05 it's fine, let me know if you need help :) 13:44:42 next topic :P 13:44:53 #topic CompNeuro lab compose status check for F33/F34 13:45:00 #link https://koji.fedoraproject.org/koji/packageinfo?packageID=30691 13:45:59 There was a compose failure recently, but it was because of some low level dep (don't remember the name) 13:46:07 it got fixed, and everything's smooth 13:46:29 yeah looks fine 13:48:03 next topic ... 13:48:06 #topic Neuroscience query of the week 13:49:15 We have a query? or I need to find one? 13:49:38 Maybe FranciscoD could help us with it 13:49:40 :P 13:49:49 :P 13:49:56 a query? 13:50:04 hrm, well, nothing off the top of my hea 13:50:06 *head 13:50:47 what are folks interested in? 13:51:06 there's this paper on how deep learning informs neuroscience: https://www.nature.com/articles/s41583-020-00395-8 13:51:30 I have access to it, so if someone wants the pdf, let me know :) 13:51:51 (if it isn't open access, that is) 13:51:56 ;) 13:52:21 I'll add it to the reading list too---worth a skim 13:52:25 it isn't open access... will have to connect to my institute's network 13:52:35 My university has subscription too 13:52:35 pagure.issue.comment.added -- ankursinha commented on ticket neuro-sig/NeuroFedora#318: "NeuroFedora reading/listening list" https://pagure.io/neuro-sig/NeuroFedora/issue/318#comment-701656 13:52:54 let me know if you can't get it, I'll e-mail it to you 13:53:18 cool, thanks FranciscoD++ 13:53:18 bt0: Karma for ankursinha changed to 4 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 13:53:37 #topic Next meeting day, and chair. 13:53:37 Isn't the current understanding of perception of Neural nets similar to that of the brain? 13:53:46 I always thought it was 13:53:46 #undo 13:53:46 Removing item from minutes: 13:53:58 MeWjOr: what do you mean by "perception"? 13:54:15 umm... interesting XD 13:54:27 neural nets do specific tasks---they can be trained to do specific ones 13:55:06 and there are similarities with the brain, but neural nets don't generalise etc---and online learning is still not easy 13:55:21 basically I meant how CNNs understand the patterns in the fourier domain, so in that we defined the "perception" from the filter size 13:55:30 sort of like, how small features it can understand 13:55:45 well, the question is whether it is actually "understanding" anything :D 13:56:20 not understanding, but recognising the pattern... 13:56:30 but yes, not understanding 13:56:36 Yeh, that's a better word I guess 13:56:48 sounds better 13:57:17 but since they're rate based, i.e., rates of neuronal firing is what is used, they ignore the information that the brain has in "spikes" 13:57:34 and the timing between spikes (inter-spike intervals) etc. 13:57:59 If you want to see how RNNs are used in neuroscience, look for David Sussillo's papers. They've been using them quite a bit recently 13:58:22 cool 13:58:28 https://www.sciencedirect.com/science/article/pii/S0896627309005479 13:58:31 next topic 13:58:45 Are people still working on RNNs for neuroscience? Isn't the transformer network the new thing now? 13:58:56 https://www.mitpressjournals.org/doi/10.1162/NECO_a_00409 13:59:10 idk, machine learning stuff moves really fast XD 13:59:16 no, RNNs are still around---with chaotic etc. 13:59:28 but it depends on whether one is trying to do ML, or if one is trying to understand the brain 13:59:40 ohh, yes 14:00:05 thanks for the references... I'll try to look them up :D 14:00:13 The real science don't understand about fashions 14:00:25 #topic Next meeting day, and chair 14:01:28 2 weeks again? seems to work well 14:01:33 MeWjOr or I can chair? 14:01:38 we need a chair for the next meeting in November 30, some one is a volunteer? 14:01:47 That would be my end-sem test month... I don't think I will be able to chair 14:01:52 or me again, This hour is good for me 14:01:55 I'll chair, no worries 14:02:00 Great 14:02:01 :D 14:02:04 bt0: good for us all to take turns :) 14:02:06 cool 14:03:19 #action Next meeting November 30, same time in two weeks. FranciscoD will chair the meeting 14:03:31 #topic openfloor 14:03:45 #action bt0 send out the meeting logs 14:04:02 Anything for Open Floor? 14:04:30 nothing from me 14:04:42 will look into packages---busy times at the end of the year, unfortunately 14:05:47 yeah 14:06:12 the whole project slows down theri activities bc the holidays 14:06:22 *their 14:06:31 I will close the meeting 14:07:10 #endmeeting