[17:15:18] <leila>	 bmansurov: shall I put the instruction in Notes tab for Section translation spreadsheet?
[17:15:35] <bmansurov>	 leila: yes, that'd be nice.
[17:15:42] <leila>	 ok. working on it.
[17:15:48] <bmansurov>	 thanks
[19:16:40] <halfak>	 o/
[19:16:46] <halfak>	 What's the stream link for the showcase?
[19:16:54] <halfak>	 bmansurov, need help changing the topic?
[19:17:17] <bmansurov>	 halfak, sure, I'll paste the link here.
[19:17:49] <bmansurov>	 halfak https://www.youtube.com/playlist?list=PLhV3K_DS5YfLQLgwU3oDFiGaU3K7pUVoW
[19:18:09] <halfak>	 bmansurov, I don't think that is right. 
[19:18:35] <halfak>	 https://www.youtube.com/watch?v=fpmRWCE7F_I
[19:18:41] <bmansurov>	 halfak, oops https://www.youtube.com/watch?v=fpmRWCE7F_I
[19:18:46] <bmansurov>	 yes!
[19:20:41] <halfak>	 Showcase starts in 10 minutes!
[19:21:34] <halfak>	 Nice
[19:23:51] <halfak>	 bmansurov, you must keep the note about public logging in there
[19:24:04] <bmansurov>	 ok, a sec
[19:27:00] <halfak>	 Starting in 4 minutes!
[19:31:28] <bmansurov>	 Hello, everyone. Please send your questions to me and I'll read them out.
[19:33:18] <dsaez>	 there is some background noise
[19:34:34] <leila>	 dsaez: yup, you mean in Dario's room? 
[19:34:41] <brendan_campbell>	 it was me, should be gone now
[19:34:47] <dsaez>	 maybe
[19:34:48] <leila>	 brendan_campbell: ok. thanks.
[19:34:48] <DarTar>	 yo people
[19:34:57] <leila>	 yooo DarTar
[19:37:26] <Giovanni_Ciampag>	 Hola
[19:37:36] <leila>	 ow hey Giovanni_Ciampag. :)
[19:37:52] <halfak>	 o/
[19:46:08] <halfak>	 Uh oh.  Graph doesn't start at zero 
[19:46:14] <halfak>	 That's a demerit. :P 
[19:48:26] <dsaez>	 hehe
[19:52:22] <halfak>	 bmansurov, question for Miriam.  What kind of processing power is necessary for running the kind of image processing algorithms we really want to use to sort through commons?  What kind of processing resources should we be investing in now?
[19:52:42] <bmansurov>	 ok, halfak
[19:56:04] <Giovanni_Ciampag>	 bmansurov: question for Miriam, what weights did the learning-to-rank model find when you trained it? Did it give more power to quality or to relatedness?
[19:56:34] <bmansurov>	 ok, Giovanni_Ciampag 
[19:58:12] <DarTar>	 question for miriam: what have you learned about the two specific wikidata classes you’ve worked with and why were they selected in the first place?
[19:59:31] <Giovanni_Ciampag>	 Love wikishootme, so cool!
[20:00:07] <halfak>	 Great!
[20:01:38] <Giovanni_Ciampag>	 Thanks!
[20:08:30] <Nettrom>	 I checked SuggestBot’s configuration, we recommend orphan articles to users in English, Russian, Persian, and French
[20:08:43] <Nettrom>	 also, hi everyone :)
[20:15:11] <DarTar>	 HOAX
[20:15:32] <Giovanni_Ciampag>	 GOOD
[20:15:37] <DarTar>	 HOAX
[20:15:42] <Giovanni_Ciampag>	 D'oh!
[20:15:51] <DarTar>	 gee I’m good at this
[20:15:58] <Giovanni_Ciampag>	 ehehehe
[20:16:09] <HaeB>	 hangouts obscures the crucial last lines (where the citation should be ;)
[20:16:21] <brion>	 citation neeeeeeded :D
[20:16:26] <HaeB>	 good
[20:23:32] <DarTar>	 WikiProjects are sooo good as topic labels
[20:23:43] <harej>	 ([[Wikipedia:WikiProject Directory]] is better.)
[20:25:03] <DarTar>	 https://twitter.com/wpstubs still using parsed wikiprojects names as hashtags every minute
[20:28:41] <HaeB>	 question for halfak:  a system like this is already in place on English Wikipedia with Article Alerts and NewArtBot (e.g. new articles of interest to WikiProject Morocco: https://en.wikipedia.org/wiki/User:AlexNewArtBot/MoroccoSearchResult ) - based on a simple search term list  ... 
[20:28:46] <HaeB>	  ... did we evaluate the improvements we gain from the new more sophisticated ML approach?
[20:29:09] <bmansurov>	 ok
[20:31:09] <DarTar>	 clarification question for halfak: the draft topic model is really an article topic classifier, nothing specific to drafts, correct? And what are your thoughts about cross-language generalization issues?
[20:39:08] <HaeB>	 this is the search regex list BTW for that example: https://en.wikipedia.org/wiki/User:AlexNewArtBot/Morocco
[20:40:31] <Giovanni_Ciampag>	 Thanks Miriam and Aaron, very interesting talks!
[20:40:31] <Nettrom>	 great showcase, thanks!
[20:40:37] <halfak>	 HaeB, yeah, so I think the automated nature of the OREs model will be more useful for this. 
[20:40:49] <halfak>	 But maybe, this will help with being more fine-grained. 
[20:41:13] <DarTar>	 thanks all 
[20:41:20] <halfak>	 E.g. you might imagine a mixed filter of "geo.countries" and these regexes to build a list. 
[23:31:59] <HaeB>	 J-Mo: https://wikitech.wikimedia.org/w/index.php?title=PAWS&diff=prev&oldid=1783397 https://www.mediawiki.org/w/index.php?title=Manual:Pywikibot/PAWS&diff=prev&oldid=2720960 ;)
[23:57:59] <dsaez>	 hey halfak, qq: I'm playing with ORES and I don't get a probability for A category, am I doing something wrong or it works like this ? (in English) 
[23:58:11] <halfak>	 what?
[23:58:17] <dsaez>	 predictions
[23:58:21] <halfak>	 A as in a quality class?
[23:58:23] <halfak>	 What model?
[23:58:33] <dsaez>	 wp10
[23:58:40] <halfak>	 "A" class doesn't exist 
[23:58:46] <halfak>	 Or, well it's an idea but not a reality. 
[23:58:51] <halfak>	 No one ever used it.
[23:58:55] <halfak>	 So we don't learn it. 
[23:59:08] <dsaez>	 ooh, I see. 
[23:59:33] <dsaez>	 thanks (I'm having a lot of fun playing playing with ORES)
[23:59:45] <dsaez>	 amazing tool.