[18:57:55] * guillom does the halfak summoning dance. [19:07:24] It worked! [19:09:20] HaeB / halfak|Lunch : I'm looking for a study (or quote from a study) that highlights how much our content is used to train AI-related tools & research. I know that's the case (from reading many of those papers) but I'm in a bit of a time crunch and I'd really need to find a solid reference. [19:10:12] guillom, I know that Brent Hecht and Shilad Sen like to talk about the use of Wikipedia in algorithm research [19:10:23] You might beat me by checking google scholar for that now. [19:10:34] Otherwise, I'll do some searching in 15 minutes [19:11:09] halfak|Lunch: Thank you! I'd appreciate the help :) [19:13:02] "Shilad" is in here from time to time [19:20:57] guillom, "Behind the scenes, many important [19:20:57] intelligent algorithms utilize data from Wikipedia and OSM [19:20:57] to make geographic inferences about the world [20,27,44]. " [19:21:06] Johnson, I. L., Lin, Y., Li, T. J. J., Hall, A., Halfaker, A., Schöning, J., & Hecht, B. (2016, May). Not at home on the range: Peer production and the urban/rural divide. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 13-25). ACM. [19:21:50] halfak: <3 Thank you! [19:22:01] No problem :) [19:45:51] 10Quarry: SQL Syntax errors in Quarry - https://phabricator.wikimedia.org/T188538#4011560 (10Tohaomg) [20:14:18] guillom: there's this recent overview paper https://meta.wikimedia.org/wiki/Research:Newsletter/2018/February#Researching_the_research_using_Wikipedia_as_a_corpus [20:14:55] haven't read it and don't know how much it highlights AI/ML applications specifically, but a lot of the NLP stuff is based on ML [20:15:47] (we honestly skip much of that for the research newsletter, because it's not primarily about insights about WP itself, more often about training/testing algorithms) [20:22:05] 10Quarry: Implement SQL Query Validator in Quarry - https://phabricator.wikimedia.org/T188538#4011705 (10Reedy) [20:48:56] halfak: whenever you get a chance can you process https://phabricator.wikimedia.org/project/board/45/query/2y.A4L2AFfDN/ ? I'm trying to make sense of what needs processing on Research board. ;) [20:50:44] lzia, processed [20:50:47] :) [20:51:01] vaa! you're great halfak. :D [20:51:06] thanks. I will sleep better tonight. [20:51:17] You better :P [20:51:21] :P [20:51:28] let me try it with someone else then. [20:52:36] J-Mo: can you check to see if https://phabricator.wikimedia.org/project/board/45/query/AIDuHJyVvpAc/ reflects what it should reflect? :D [20:54:57] lzia: thanks for keeping me accountable. I moved the GLAM SDC task to 'done'. Three of the in progress tasks are current; the fourth is… in indefinite limbo and I need to reach out to the External Collaborator to sunset the project, but have been putting that off. Realistically, it's done on my part, but wrap up feels like a huge hassle so I've been hiding from it [20:55:40] thanks, J-Mo. anything I can help with the last one? [20:56:39] J-Mo, maybe we should move that to blocked until you're ready to pick it up. [20:57:52] lzia: let's block it, yes. I owe Amy Z and email; no excuse for not sending it other than anxiety/laziness. I'll add it to my Monday AM tasks and try not to ignore it again [20:58:45] thanks, J-Mo. and btw, I'm thinking of the Phab as a place we can see who's working on what, beyond what we report on the Monday meeting. so my interest in having it clean is related to that. ;) [21:00:38] agree that would be useful. I don't generally use Phab that way, but I will try to be more disciplined [21:24:00] 6th birthday https://twitter.com/WikiResearch/status/968949260254199809 [22:52:46] 10Quarry: Implement SQL Query Validator in Quarry - https://phabricator.wikimedia.org/T188538#4011560 (10zhuyifei1999) Quarry isn't supposed to be slow in 'queued'. I'll investigate. [23:05:11] 10Quarry: Quarry should refuse to save results that are way too large - https://phabricator.wikimedia.org/T188564#4012234 (10zhuyifei1999) [23:09:28] 10Quarry: Quarry should refuse to save results that are way too large - https://phabricator.wikimedia.org/T188564#4012265 (10zhuyifei1999) https://www.mediawiki.org/wiki/Topic:U1le4hrq6eunlafz says 500k works. NO, that's waaay tooo large, you should be using dumps if you want so much unfiltered data. Gonna limit... [23:09:55] 10Quarry: Quarry should refuse to save results that are way too large - https://phabricator.wikimedia.org/T188564#4012267 (10zhuyifei1999) p:05Triage>03High