[07:57:52] 10Scoring-platform-team-Backlog, 10editquality-modeling, 10revscoring, 10Easy, 10artificial-intelligence: Progress indicator for editquality autolabel utility - https://phabricator.wikimedia.org/T172225#3491077 (10Ladsgroup) https://github.com/wiki-ai/editquality/pull/91 [07:58:22] 10Scoring-platform-team, 10editquality-modeling, 10revscoring, 10Easy, and 2 others: Progress indicator for editquality autolabel utility - https://phabricator.wikimedia.org/T172225#3564991 (10Ladsgroup) a:03Ladsgroup [08:30:52] 10Scoring-platform-team, 10Bad-Words-Detection-System, 10revscoring, 10Malayalam-Sites, and 2 others: Add language support for Malayalam - https://phabricator.wikimedia.org/T173193#3565029 (10Ladsgroup) The script is running now, the result will be in https://meta.wikimedia.org/wiki/Research:Revision_scori... [08:33:35] 10Scoring-platform-team, 10Bad-Words-Detection-System, 10revscoring, 10User-Ladsgroup, 10artificial-intelligence: Add language support for Urdu - https://phabricator.wikimedia.org/T173190#3565030 (10Ladsgroup) It doesn't have the informal words :( [09:12:31] 10Scoring-platform-team, 10Bad-Words-Detection-System, 10revscoring, 10User-Ladsgroup, 10artificial-intelligence: Add language support for Croatian (hr.wiki) - https://phabricator.wikimedia.org/T172046#3565117 (10Ladsgroup) [09:13:29] 10Scoring-platform-team, 10Bad-Words-Detection-System, 10revscoring, 10User-Ladsgroup, 10artificial-intelligence: Add language support for Croatian (hr.wiki) - https://phabricator.wikimedia.org/T172046#3483983 (10Ladsgroup) https://github.com/wiki-ai/revscoring/pull/352 [09:14:56] wiki-ai/revscoring#1201 (hrwiki - b934402 : Amir Sarabadani): The build failed. https://travis-ci.org/wiki-ai/revscoring/builds/269911637 [09:19:06] wiki-ai/revscoring#1203 (hrwiki - dec23b1 : Amir Sarabadani): The build was fixed. https://travis-ci.org/wiki-ai/revscoring/builds/269912875 [09:22:09] 10Scoring-platform-team-Backlog, 10Wikilabels, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Edit quality campaign for hr.wiki - https://phabricator.wikimedia.org/T172047#3565125 (10Ladsgroup) @Ivi104: Hey, thank you for working on this. Some stuff needed to move on with this. Firs... [09:35:36] 10Scoring-platform-team-Backlog, 10Bad-Words-Detection-System, 10revscoring, 10User-Ladsgroup, 10artificial-intelligence: Add language support for Bosnian - https://phabricator.wikimedia.org/T173087#3565150 (10Ladsgroup) It's running now, it'll be done in less than six hours. Tell me when it's not there. [11:11:47] 10Scoring-platform-team, 10Operations, 10Ops-Access-Requests, 10Graphite, 10User-fgiunchedi: Temporarily access request to graphite nodes - https://phabricator.wikimedia.org/T174542#3565331 (10Ladsgroup) [11:19:06] 10Scoring-platform-team, 10DBA, 10Operations, 10cloud-services-team: Labsdb* servers need to be rebooted - https://phabricator.wikimedia.org/T168584#3369055 (10faidon) JFTR, since I didn't see it mentioned neither here nor in T142807, how impending is that decomm? Days/weeks/months? [11:51:03] 10Scoring-platform-team, 10Operations, 10Ops-Access-Requests, 10Graphite, 10User-fgiunchedi: Temporarily access request to graphite nodes - https://phabricator.wikimedia.org/T174542#3565331 (10MarcoAurelio) Amir is trusted, support from me. [12:49:49] 10Scoring-platform-team-Backlog, 10Wikilabels, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Edit quality campaign for hr.wiki - https://phabricator.wikimedia.org/T172047#3565529 (10Ivi104) @Ladsgroup The messages have been translated! The phrase "Edit quality (20k sample)" transla... [13:32:55] 10Scoring-platform-team, 10Operations, 10Ops-Access-Requests, 10Graphite, 10User-fgiunchedi: Temporarily access request to graphite nodes - https://phabricator.wikimedia.org/T174542#3565694 (10Addshore) [13:39:28] o/ [13:39:36] UI is going crazy. Going for a reboot :/ [13:42:22] 10Scoring-platform-team, 10Operations, 10Ops-Access-Requests, 10Graphite, 10User-fgiunchedi: Temporarily access request to graphite nodes - https://phabricator.wikimedia.org/T174542#3565714 (10Ladsgroup) I've got list of them so I don't need this anymore. Thanks :) [13:42:27] 10Scoring-platform-team-Backlog, 10ORES, 10Operations, 10Graphite, 10User-fgiunchedi: Regularly purge old ores graphite metrics - https://phabricator.wikimedia.org/T169969#3565716 (10Ladsgroup) [13:42:29] 10Scoring-platform-team, 10Operations, 10Ops-Access-Requests, 10Graphite, 10User-fgiunchedi: Temporarily access request to graphite nodes - https://phabricator.wikimedia.org/T174542#3565715 (10Ladsgroup) 05Open>03Resolved [13:44:49] halfak: hey, I have lots of things to talk to you [13:45:17] https://phabricator.wikimedia.org/T169969 can you give me list of things that can be removed after a month [13:46:02] Amir1, thanks for picking that up. [13:46:08] BTW, the svwiki campaign just finished. [13:46:10] \o/ [13:46:31] It seems I have a huge undertime for ORES, I will work full time today and tomorrow on ores only \o/ [13:46:47] Nice, I need things to do [13:47:00] more ores is better ores :D [13:47:23] halfak: there are several PRs waiting for you in case you didn't know [13:48:12] Link me in priority order and I'll try to clear 'em out right now. [13:51:18] https://github.com/wiki-ai/revscoring/pull/352 [13:51:22] https://github.com/wiki-ai/editquality/pull/91 [13:53:29] 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete svwiki edit quality campaign - https://phabricator.wikimedia.org/T174557#3565764 (10Halfak) [13:53:47] also https://phabricator.wikimedia.org/T169969 is blocked on you for now :D [13:53:48] 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete svwiki edit quality campaign - https://phabricator.wikimedia.org/T174557#3565778 (10Halfak) http://labels.wmflabs.org/stats/svwiki/ Looks like we're done. [13:56:47] 10Scoring-platform-team, 10ORES, 10editquality-modeling, 10artificial-intelligence: Deploy damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174558#3565785 (10Halfak) [13:57:18] 10Scoring-platform-team, 10ORES, 10editquality-modeling, 10artificial-intelligence: Train & test damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174559#3565798 (10Halfak) [13:57:34] 10Scoring-platform-team, 10ORES, 10editquality-modeling, 10artificial-intelligence: Train & test damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174559#3565798 (10Halfak) [13:57:35] 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete svwiki edit quality campaign - https://phabricator.wikimedia.org/T174557#3565811 (10Halfak) [13:58:58] 10Scoring-platform-team-Backlog, 10Collaboration-Team-Triage, 10ORES, 10editquality-modeling, 10artificial-intelligence: Enable ORES filters for svwiki - https://phabricator.wikimedia.org/T174560#3565822 (10Halfak) [13:59:15] 10Scoring-platform-team, 10ORES, 10editquality-modeling, 10artificial-intelligence: Deploy damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174558#3565785 (10Halfak) [14:02:21] 10Scoring-platform-team-Backlog, 10ORES, 10editquality-modeling, 10artificial-intelligence: Deploy damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174558#3565844 (10Halfak) [14:04:37] 10Scoring-platform-team, 10draftquality-modeling, 10artificial-intelligence: [Discuss] draftquality on a sample, humongous everything, or something else? - https://phabricator.wikimedia.org/T168909#3565855 (10Halfak) ``` roc_auc (micro=0.979, macro=0.97): vandalism attack spam OK... [14:34:49] o/ Amir1 ! [14:35:22] halfak: Amir1: Is this a gnuplot shop, or something like ipython-notebook + matplotlib? [14:35:22] awight: hey! [14:35:39] ggplot2 :D [14:35:41] R [14:35:52] But I'd be interested in switching towards bokeh and jupyter [14:36:18] I'd miss data.table but pandas is OK I guess. [14:37:29] ggplot2 from within R, or standalone? [14:37:35] from within R [14:37:45] You can do R in jupyter notebooks these days. [14:37:52] that was fast :D [14:38:09] lol was reviewing that :P [14:38:14] s/ok was just about to merge too [14:38:22] Amir1: on the Makefile tip, check this out… https://github.com/wiki-ai/editquality/blob/templating/Makefile.j2 [14:38:34] We need to make reverted for hrwiki too [14:38:52] oh nice [14:39:22] It’s just a proof-of-concept, but compiles to an almost byte-identical Makefile. [14:40:27] The “fun” part is that we would have to list out all variations on the workflow and find a way to (in?)flexibly represent them in the templates [14:41:03] halfak: Do you have sample code I can crib? So far, I’ve thrown this junk together: https://github.com/adamwight/thresholds_diagrams/blob/master/draftquality_thresholds.png [14:42:10] awight, cool! I suppose we must generate from python if we're looking to build plots for thresholds. [14:42:31] I haven't used bokeh for much yet. I like it in concept because it implements the Grammar of Graphics. [14:42:51] Which removes a ton of incidental complexity from plotting. [14:43:02] ggplot2 is also based on the Grammar of Graphics. [14:43:23] R+ggplot2 ~ Python+bokeh [14:49:30] yas. [14:49:32] Change 374593 merged by Ottomata: [14:49:33] [operations/puppet@production] Let scoring platform team run "lsof" for diagnostics [15:00:33] 10Scoring-platform-team-Backlog, 10Wikilabels, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Edit quality campaign for hr.wiki - https://phabricator.wikimedia.org/T172047#3566082 (10Ladsgroup) [15:00:50] 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Edit quality campaign for hr.wiki - https://phabricator.wikimedia.org/T172047#3483996 (10Ladsgroup) [15:01:34] 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Edit quality campaign for hr.wiki - https://phabricator.wikimedia.org/T172047#3483996 (10Ladsgroup) The UI is still in English because the new i18n hasn't arrived yet, I will make a deployment tomo... [15:08:27] halfak: awight Can I deploy wikilabels? for i18n reasons [15:08:32] +1 [15:08:39] fine by me, ty! [15:08:42] I will do another one tomorrow (when hrwiki is there) [15:09:29] awight, looking at the precision like, it seems wrong in that plot. [15:09:38] lol there’s a good chance of that [15:09:42] unless maybe you're plotting for the "OK" class [15:09:46] I guess that might make sense. :) [15:09:47] ah. yes [15:10:01] I couldn’t figure out what thresholds.true was about :). at least at midnight [15:10:15] Oh! For the damaging model. [15:10:48] kk—my plan was to go through and try to just create plots for each of the examples given [15:11:09] so should grab the damaging thresholds data set fromt the API [15:11:18] +1 [15:11:25] statistics.thresholds.true :D [15:13:32] :). ok I’ll give bokeh a shot as well [15:13:53] trying to use “lsof” at the moment, it doesn’t look like the change was propagated. [15:14:02] (the sudo change) [15:15:58] 10Scoring-platform-team-Backlog, 10Wikilabels: Wikilabels translations not updated and cannot load old worksets - https://phabricator.wikimedia.org/T174343#3558487 (10Ladsgroup) Thank you for letting me know, I just deployed a new version which adds those new translations and it should be okay now. We close th... [15:16:12] 10Scoring-platform-team, 10Wikilabels, 10User-Ladsgroup: Wikilabels translations not updated and cannot load old worksets - https://phabricator.wikimedia.org/T174343#3566175 (10Ladsgroup) a:03Ladsgroup [15:17:31] is anyone doing this? https://phabricator.wikimedia.org/T174559 I'm on it [15:17:35] halfak: awight [15:18:13] Amir1: I hadn’t touched it [15:18:48] 10Scoring-platform-team, 10ORES, 10Operations, 10Graphite, 10User-fgiunchedi: Regularly purge old ores graphite metrics - https://phabricator.wikimedia.org/T169969#3566203 (10Ladsgroup) [15:25:51] 10Scoring-platform-team, 10ORES, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Train & test damaging/goodfaith model for svwiki - https://phabricator.wikimedia.org/T174559#3566248 (10Ladsgroup) a:03Ladsgroup [15:25:53] 10Scoring-platform-team, 10ORES, 10Operations: rack/setup/install ores2001-2009 - https://phabricator.wikimedia.org/T165170#3566244 (10akosiaris) 05Open>03stalled Stalling a bit while T169246 is ongoing [15:26:05] halfak: I picked it up: https://phabricator.wikimedia.org/T174559 [15:26:40] 10Scoring-platform-team, 10Bad-Words-Detection-System, 10revscoring, 10User-Ladsgroup, 10artificial-intelligence: Add language support for Croatian (hr.wiki) - https://phabricator.wikimedia.org/T172046#3566253 (10Ladsgroup) [15:26:55] cool [15:26:56] :) [15:26:58] 10Scoring-platform-team, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Reverted model for hrwiki - https://phabricator.wikimedia.org/T174572#3566257 (10Ladsgroup) [15:27:27] 10Scoring-platform-team, 10Wikilabels, 10User-Ladsgroup: Wikilabels translations not updated and cannot load old worksets - https://phabricator.wikimedia.org/T174343#3566275 (104shadoww) Thank you, looks good. As I thought that workset thing was some kind of feature:) [15:28:10] 10Scoring-platform-team, 10Wikilabels, 10User-Ladsgroup: Wikilabels translations not updated and cannot load old worksets - https://phabricator.wikimedia.org/T174343#3566279 (104shadoww) 05Open>03Resolved [15:31:12] Amir1, don't forget to back-fill that deploy task :D [15:31:30] with pleasure :D [16:02:56] halfak: still can't join :( [16:29:10] and of course I froze [16:30:09] I’m “join”ing like crazy and not getting in [16:43:01] 10Scoring-platform-team, 10DBA, 10Operations, 10cloud-services-team: Labsdb* servers need to be rebooted - https://phabricator.wikimedia.org/T168584#3369055 (10chasemp) No definite date has been set as we are working on T173511 as the precursor to moving over quarry (and probably PAWS). I think we are tal... [16:49:01] 10Scoring-platform-team, 10Edit-Review-Improvements-RC-Page, 10MediaWiki-extensions-ORES, 10Collaboration-Team-Triage (Collab-Team-Q1-Jul-Sep-2017): Very long search times on RC Page for "Very likely good faith" + "Likely have problems" (on en.wiki only?) - https://phabricator.wikimedia.org/T164796#3566723 (... [16:53:20] Amir1: Did you want to do Scrum of Scrums today? [16:53:56] I'm busy doing other things but if you need to step off, I'm game [17:06:14] Amir1: no worries, I enjoy it! [17:12:45] 10Scoring-platform-team-Backlog, 10Wikilabels, 10articlequality-modeling, 10artificial-intelligence: Build article quality model for euwiki - https://phabricator.wikimedia.org/T171119#3566857 (10Theklan) Hello! Any news on this? We will start soon with the Thanks! [18:33:54] Tragic! Interactive Bokeh images don’t render in Github: https://github.com/adamwight/thresholds_diagrams/blob/master/Thresholds%20diagrams.ipynb [18:35:45] Looks like a “won’t fix”, https://github.com/bokeh/bokeh/issues/4518 [18:37:25] damn [18:41:02] halfak: Is this the graph you were expecting, btw? https://github.com/adamwight/thresholds_diagrams/blob/master/damaging.png [18:41:29] awight, strange. recall should go down as threshold increases. [18:41:43] Filter rate looks right :) [18:42:24] Recall must go down and precision goes up. [18:42:37] I wonder if it turns out the data just looks that way. [18:42:47] In which case, we probably have a bug in the stats. [18:43:17] halfak: Recall does go down in that graph—what exactly is strange? [18:44:06] Oh. my damn colorblind eyes [18:44:09] lolol [18:44:14] * awight changes line colors [18:44:27] green and blue are too close together even for an XXX [18:45:06] The blue isn't a problem for most color visiony folks. It [18:45:10] 's green and red. [18:45:31] +1 [18:50:25] halfak: How is this? https://snag.gy/TdCAKh.jpg [18:50:50] Much better! [18:51:10] Looks like an interesting bit about precision is covered by the legend though. [18:51:18] cool—fixing [19:07:21] https://github.com/adamwight/thresholds_diagrams/blob/master/damaging.svg [19:07:55] halfak: While I have the framework running, it would be easy to make more plots if it might be useful for your blog post? [19:12:48] awight, oh the dots are awesome! [19:13:20] Can you make one that plots !precision and !recall for the "OK" class of draftquality? [19:13:43] I have an updated model. Are you pulling this from ores-misc now? [19:13:57] no, lemme run the makefile again :) [19:14:13] Cool. I'll submit my new model PR. [19:14:25] I've been holding off because it's kind of big :/ [19:14:29] 75MB --> 125MB [19:15:56] every MB we commit brings us that much closer to git-lfs [19:16:01] lol [19:16:06] * halfak dumps in random noise [19:16:22] Yeah. Github won't even accept a 125MB file. [19:16:23] damn [19:16:25] hmm [19:16:39] baahahaha [19:16:55] OK trying something else. Don't block on me :\ [19:17:23] I’ll get the -OK graph up, then leave it to re-run with the new model. [19:17:51] daily meta-ores candyland intermission... [19:18:17] Mind if we just switch bad codenames to “JADE” for now? [19:18:27] I can’t say “meta-ores” too many mores [19:20:04] wrt the mining metaphor, it turns out it’s incredibly easy to find jade, but the high-quality stuff is what is sought after [19:20:04] https://en.wikipedia.org/wiki/Jade#Mining [19:20:11] perhaps that’s apt [19:20:38] “more manageable 10-tonne pieces” whew [19:20:44] +1 [19:21:20] * awight sneaks more stuff through ipso facto [19:25:29] re. jade schema, I’m drawing ever closer to the Loomio model. [19:25:50] Each wiki artifact is a statement that can be judged and discussed [19:26:16] discussion is a timeline of heterogenous data types [19:26:19] events [19:27:40] Users can make a judgment, one unit of which consists of any 1.* comments or scores on arbitrary scales. [19:28:09] "1.*"? [19:28:31] Seems to me we shouldn't re-implement discussions [19:28:38] Whenever a judgement introduces a scale on an artifact, all judgments a collection of scores that can be queried [19:28:58] +1 I’d like to attach an (opaque) Flow thread to each artifact [19:29:04] Gotcha. [19:29:08] +1 [19:29:41] I’m bothered, though, that the discussion thread is orthogonal to the scores yet should probably still reference events somehow. [19:29:56] It’s natural to back up a score with commentary. [19:30:27] So I guess, judgment is at the center of a star topology [19:30:45] A judgment points to 0..1 flow posts [19:31:33] 0..* scores point to a judgment (must have the same author) [19:31:43] Maybe comment != discussion [19:31:47] hmm [19:31:56] it should appear in the same timeline [19:32:03] could be different types of event [19:32:17] But wouldn’t we want them to show in the Flow thread? [19:32:32] * awight stops coupling to Flow [19:32:37] I don't see why that is a problem [19:32:42] OK [19:32:53] Each score/endorsement gets a comment (plain text, unicode) and optionally a discussion can be associated with each wiki-thing. [19:33:07] -1 plain text eh [19:33:16] KISS [19:33:19] YAGNI [19:33:20] :P [19:33:31] done +1 [19:33:33] Can always promise to render something later [19:33:58] I guess we could make the same guarantees as wiki comments. [19:33:58] Mmm how about we propose this and maybe someone else will wire VE in [19:33:59] Links only. [19:34:05] sure [19:34:07] A-Ok [19:34:25] I’d like to allow editing of comments and of scores [19:34:27] Drawback of having comments be part of our schema is they will need alternative patrol. [19:34:33] +1 [19:34:38] Changing endorsements. [19:34:49] wondering how events and an editable thing can be reconciled? [19:34:58] We send tombstones and change events? [19:35:14] yuck [19:35:33] changes need history [19:35:45] I think KISS in this case dictates that we reuse wiki machinery for all that [19:36:04] Judgement is a content type [19:36:15] err ^content handler? [19:37:42] How will this look if viewed as a timeline… only show updated score? group by author and show their history, default to collapsed view as most recent? [19:37:48] has loomio solved this? [19:38:01] * awight updates etherpad [19:47:25] awight, events [19:47:33] Every updated endorsement is an event [19:47:44] The history of a thing is the events relevant to that thing [19:48:01] Seems like you are asking UX questions [19:48:31] I’m thinking, it’s meaningful if a person judges twice using the same scale, and although we only tally the latter, comparing the two gives us another source of data. [19:49:01] Judges the same thing twice? [19:49:12] i.e., swinging their vote [19:49:13] Like, changes their judgement? [19:49:15] yeah [19:49:23] Gotcha. Both events would be in the event history [19:49:46] The tallying system (whatever that is) would only be concerned by the most up to date judgements. [19:49:48] The UX does matter to schema, I think, since it determines many use cases. But I guess that since both schema and UX have a dependency on designing the use cases, we should do that early as well. [19:50:19] I disagree. We should ask the question "what groupings are important" and not "what should the default grouping be" [19:50:38] One is about affordances and the other is about a specific context and experience. [19:53:50] til affordance [19:55:23] Though, to be honest, if we could get a proper designer working on a potential UI for this (context/experience) I think that would be very informative. [19:55:31] \p/ [20:07:46] hiring a designer sounds like a great idea [20:30:34] is “JADE” or “JaDE” less annoying? [20:30:53] with the latter, there’s a good chance that everyone will have pronounciation fail almost all of the time 8D [20:31:01] jah-day [20:31:11] oh res [20:31:36] mor-ee-oh ORES-ee-ohs [20:32:07] I think we may have found two l10n-proof words [20:32:40] who from a phonetic language would think that “jade” is pronounced rhyming with “stayed” [21:07:20] https://github.com/adamwight/thresholds_diagrams/blob/master/draftquality-OK.svg [21:07:23] Does it need any markers? [21:10:35] and did you find a way around the 125MB GH limit? [21:10:47] 24 more hours of modeling…? [21:11:49] * awight rubs hands delightedly at thought of enjoying Amir1’s progress indicator during my next bout of model-building purgatory. [21:12:08] :D [21:14:18] I would not mind seeing [[File:Precisionrecall.svg|thumb|Precision and recall]] on every page that references P-R... [21:14:19] 10[1] 04https://meta.wikimedia.org/wiki/File:Precisionrecall.svg [21:14:35] AsimovBot: You did it, you won my heart! [21:14:35] 04Error: Command “you” not recognized. Please review and correct what you’ve written. [21:14:45] Despite everything [21:15:36] slightly glitchy chatbots might be slightly more futuristic than clean ones [21:19:48] halfak: Do any of our grantmakers want to be publicly acknowledges as stakeholders? [21:20:22] awight, wat [21:20:31] *acknowledged [21:20:52] Off to an esoteric start, https://www.mediawiki.org/wiki/ORES/Stakeholders [21:21:12] Seems nice, as well as transparent, to list our grantmakers here [21:21:18] Oh. [21:21:42] Who is an example "grantmaker" [21:21:44] ? [21:22:16] Also, tool devs should be independently important from end-users. [21:22:21] yes [21:22:27] sorry, just getting started. [21:22:43] no worries [21:22:48] just having thoughts ^_^ [21:22:52] I believe grantmaker is what you call people who give us grants. [21:23:10] open to a less jargony name [21:24:12] So, like WMF Grants: Project grants? [21:24:21] this is interesting https://www.bsr.org/reports/BSR_Stakeholder_Engagement_Stakeholder_Mapping.final.pdf [21:24:22] or like the Wikimedia Foundation who pays our salaries. [21:24:35] k yeah I have WMF, but project grants are a separate thing eh [21:24:49] Gotcha. yeah. [21:24:49] just as we in scoring team are separate from the org [21:25:03] Se have researchers who are going to be including judgements in their studies. [21:25:24] k found the cookie crumbs, https://meta.wikimedia.org/wiki/Grants:IEG/Revision_scoring_as_a_service/Renewal [21:25:29] yes! [21:25:37] the more diverse the stakeholders, the better. [21:34:34] Am I being a jerk to include dates for funding sources? [21:34:59] I don't think so [21:35:42] Amir1: Is it kosher to say you are “sponsored” by WMDE? I know there’s some German law about saying “employee”, but unsure how strict that is. See [[mw:ORES/Stakeholders]] [21:35:44] 10[2] 10https://www.mediawiki.org/wiki/ORES/Stakeholders [21:35:51] AsimovBot: /5 [21:36:11] * awight cringes. I might have hurt a gear [21:36:51] {{done}} [21:36:52] How efficient, halfak! [21:36:55] <3 [21:38:27] {{inline-code | phpinfo(); }} foo [21:38:27] 10[3] 04https://meta.wikimedia.org/wiki/Template:inline%2Dcode [21:38:36] lol [21:39:00] {{mw:inline-code | phpinfo(); }} [21:39:01] 10[4] 04https://meta.wikimedia.org/wiki/Template:mw:inline%2Dcode [21:39:04] I can’t watch [21:39:20] AsimovBot: you should visit more often :D [21:39:20] 04Error: Command “you” not recognized. Please review and correct what you’ve written. [21:43:05] awight: I don't know [21:43:12] I guess it's okay [21:43:31] d’oh! I’ll leave you in charge of that, sorry for the awkwardness. [22:18:41] Something we missed in today’s staff meeting. Next Monday is a holiday, but we have 4 meetings scheduled. [22:22:08] woops! [22:22:31] A bias detector: https://www.umass.edu/newsoffice/article/umass-amherst-computer-scientists-develop [22:24:22] 10Scoring-platform-team-Backlog: Use Themis bias detector on edit quality and draft quality models - https://phabricator.wikimedia.org/T174629#3568361 (10awight) [22:28:10] halfak: Can you give me +w on https://calendar.google.com/calendar/render#eventpage_6%7Ceid-NjBxbWFkaHBjNHMzNmJiNDcwcjM4YjlrNmNybWNiYjI2MHIzNmI5azc1aWppcDlvNjVoajhjYjY2MF8yMDE3MDkwNFQxNjAwMDBaIGFoYWxmYWtlckB3aWtpbWVkaWEub3Jn-0-0- [22:28:33] There’s no way that URL will go through :). That’s the “Scoring Platform sync” meeting [22:29:16] {{done}} [22:29:16] You rule, halfak! [22:29:33] Anyway, I was thinking of canceling BS time and moving the sync to 30 minutes before tech management? [22:31:29] boldly done, please revert or whatever [22:32:27] > Themis also found that designing the software to avoid discrimination against one attribute may increase discrimination against others. For example, the same decision-tree-based software trained not to discriminate on gender discriminated against race 38 percent of the time [23:09:39] halfak: Argh—the Keilana effect blog post doesn’t link to your paper, AFAICT [23:35:07] halfak: This is crappy, please tell me you already have a list somewhere so I can delete: https://www.mediawiki.org/wiki/ORES/Scholarship [23:35:32] * awight grumbles about Citoid being unavailable on mediawikiwiki [23:51:57] request to have it deployed?