[16:01:11] o/ [16:01:24] Looks like I'm not getting invited to the research showcase anymore. [16:01:31] Who should I bother about that? [16:01:40] People are scheduling over the event in my calendar :| [19:21:11] hi everyone! we're starting the Research Showcase in about 10 minutes. Today's speaker will be Leila Zia from Wikimedia Research. YouTube stream and speaker names & abstracts are linked in the Channel topic. I'll be fielding questions for the speakers during the showcase, so if you have a question, please ping me so I see it & and I'll ask it for you during the Q&A period. [19:23:11] tizianop: are you ready? :d [19:23:42] Excited. Thanks J-Mo and Leila [19:25:26] leila: yes \o/ [19:25:33] yo people [19:25:39] so looking forward to it!! [19:27:00] tizianop: do you want to join the hangout? [19:27:10] computermacgyver: it will be boring for you. reduce your expectations. ;) [19:29:13] hey Nettrom [19:29:21] hey DarTar! [19:30:06] +leila: Non-sense. (but no pressure, the work is great and speaks for itself) [19:30:29] ow the work is great. my ability to do justice is questionable. ;) [19:30:48] Arg... Missed the event because I don't have an invite. [19:30:55] Looks like I'll be able to join for talk #2 [19:30:58] we're almost ready… [19:31:20] +halfak: There is only one talk, I think [19:31:28] Bummer. [19:31:36] We are just about to start [19:31:52] alright, we're starting now! stream should be live within 30 sec [19:32:10] See you [19:32:16] It's live [19:32:24] I'm IRC host, so ping me with questions [20:15:06] As WIkidata is growing we will have good external references in Wikidata that should be availble for the reader of WIkipedia. I suggest that we have a check that all "good "external refs are in the article (I guess as a template) [20:15:40] amen to that, Magnus__ [20:16:30] Excited by Wikidata, but "good" references are sometimes language specific. enwiki prefers en-lang refs to other langs, for instance [20:16:54] (or at least I have been told this by some editors when trying to include non-english references) [20:17:14] we're wrapping up. any more questions for Leila? [20:17:35] Great work! To what extent do the related articles from category/text recommendation align with WikiProjects? Could WikiProjects be useful input/training data for any step? [20:17:46] Magnus__ would you like me to relate your comment to Leila? [20:18:04] Similar quesiton could be asked about WikiData per Magnus__'s comment [20:18:15] I have some thoughts, comments, and questions… Two thoughts: 1) Consider a WikiProject as a point of entry. 2) For English, French, and Russian, using ORES’ parameter injection can allow you to predict quality change based on adding a section of k bytes. [20:18:28] got it, computermacgyver and Nettrom [20:18:40] Injection of K bytes and 1 section :) [20:19:07] * halfak re-reads Nettrom's comment and sees that was implied. [20:19:12] what approach might they use to find and merge semantically related section titles? [20:19:21] got it, apergos [20:20:00] Q1: Maybe I missed this, but what are the reasons behind using section recommendation as the starting point for recommending to a user that they should add content to an article, as compared to other alternatives? [20:20:12] Great. Thank you. [20:20:15] BTW: WikiProjects dataset for training topic models: https://figshare.com/articles/English_Wikipedia_labeled_mid-level_wikiprojects_set/5640526 [20:20:25] Q2: Onboarding newcomers was discussed, what’s the idea behind using this to onboard newcomers? [20:20:29] And the dataset for topic spaces: https://figshare.com/articles/WikiProjects_Machine_Readable_Dataset/5503819 [20:20:36] +halfak: Thanks [20:20:36] And also, interesting project and cool work, of course :) [20:20:37] (enwiki only as leila said) [20:20:38] Leila seeks a clarification of your second thought, Nettrom [20:21:28] * DarTar adds himself to the queue [20:21:58] got your questions, Nettrom [20:21:59] halfak: oh cool, I had no idea that dataset was published [20:22:13] DarTar, was in tech mgmt ;) [20:22:36] But yeah, didn't write an announcement yet. I want to have Sumit write something up to go along with the model he is training :) [20:22:41] halfak: yeah right [20:22:49] Will loop you in DarTar :D [20:22:56] neat [20:23:58] J-Mo1: You can ask ORES what values it used for predicting the quality of a given revision, and then alter those so that it reflects that a section has been added and it will give you an updated prediction. In other words, you can use it to tell you what its prediction will be if the article was altered in a given way. Since they’re suggesting adding sections, this means you can use ORES to figure out how the predicted quality changes if a [20:23:58] user follows their suggestions. [20:24:12] any more questions for Leila? [20:24:38] <3 Nettrom's explanation [20:24:47] I hope that clarifies my thought about using ORES where its quality model is available, J-Mo1 :) [20:25:22] got it, Nettrom. If we have time after DarTar's question, I'll relate your explanation [20:25:27] thanks :) [20:26:35] thanks [20:27:00] And thanks to leila for answering my questions :) [20:29:20] +dsaez: I think Lelia said you were leading on the inter-language recommendation aspects of the project. I do some cross-language work and would be happy to chat at any point if helpful. I think inter-language section recommendations are a challenging but potentially very powerful approach [20:30:08] Wikidata can also be away to detect mismatches.... [20:32:39] computermacgiver: sure! drop me an email [20:33:04] Will do. Great to have you on the Wikimedia Research team now [20:33:19] thanks! diego at wikimedia.org :) [20:35:08] free JADE pitch [20:36:46] Happy Holidays y'all [20:37:06] Oh no the feed cut out mid- talking about upcoming stuff :( [20:37:08] Thanks everyone, great presentation and Q&A! [20:37:15] Thanks. Likewise to you [20:37:20] DarTar, what was all that you just listed off? [20:37:37] Yan Chen on Jan 17th to talk about expert engagement. Then .... [20:37:47] then second speaker tbc ;) [20:37:55] we have more than one candidate for that slot [20:38:11] cool [20:39:14] Thanks everyone. It was great to hear more about this work; very exciting [20:40:16] computermacgyver, Nettrom: thanks for the questions. [20:40:23] thanks computermacgyver tizianop Nettrom apergos and all the lurkers for joining [20:40:51] thanks for organizing these! [20:43:02] going back to my regular haunts, happy trails all [20:45:26] grabbing lunch here o/