WikiResearchVerified account

@WikiResearch

Curated research news, hot off the press. For full coverage subscribe to our newsletter. Tweets by &

San Francisco
meta.wikimedia.org/wiki/Research:…
Joined December 2011
1,142 Photos and videos Photos and videos

Tweets

You blocked @WikiResearch

Are you sure you want to view these Tweets? Viewing Tweets won't unblock @WikiResearch

  1. Pinned Tweet
    Sep 4

    In the new issue of our monthly newsletter: • Detecting spam, and pages to protect • Editors' intelligence test scores related to article quality • Three papers on Wikipedia citations

    Image
    1 reply 5 retweets 16 likes
    Undo
  2. 8 hours ago

    "Covid-on-the-Web: Knowledge Graph and Services to Advance COVID-19 Research" a dataset comprising two main knowledge graphs, including named entities linked to , and other vocabularies. (Michel et al, 2020)

    Image
    , , and 3 others
    11 retweets 20 likes
    Undo
  3. Retweeted
    10 hours ago

    Submitted my video " in the ecosystem". Had a version of 12:07 min length, then redid because they wanted 12 min max. Ended up a bit hasty at 11:44. A copy of it and the notebook sit at .

    2 retweets 4 likes
    Undo
  4. Retweeted
    Sep 18
    Replying to

    I would say so! There's some recent examples of GNNs being quite useful for mining Wikipedia, for example. See from for one example:

    1 retweet 1 like
    Undo
  5. Retweeted
    19 hours ago

    . launched the ‘Cooking with & KBpedia‘ . KBpedia intertwines seven public KGs: OpenCyc

    Image
    8 retweets 16 likes
    Undo
  6. Retweeted
    Sep 14

    Note for people using Wikipedia Cirrus dumps: some of the dumps are missing articles and it's not clear why. Most are fine, but be careful you didn't get a bad one.

    1 retweet 3 likes
    Undo
  7. Retweeted
    Sep 15

    I know a lot of people in use Attardi's wikiextractor to get raw text out of Wikipedia, but I'd recommend using Cirrus dumps instead. They come with data like categories and redirects pre-processed and plaintext is included.

    1 reply 3 retweets 5 likes
    Show this thread
    Image
    Show this thread
    Undo
  8. Sep 18

    "What if we had no Wikipedia? Domain-independent Term Extraction from a Large News Corpus" identifying “wiki-worthy” terms in a massive news corpus, with minimal dependency on actual Wikipedia entries. (Bilu et al, 2020)

    Image
    , , and
    1 reply 3 retweets 27 likes
    Undo
  9. Sep 16

    "Optimizing Hyper-Phrase Queries" spotting facts from in document collections amounting to more than thirty million documents. (Gupta and Berberich, 2020)

    Image
    , and
    3 retweets 6 likes
    Undo
  10. Retweeted
    Sep 16

    New Houcemeddine Turki, Mohamed Ali Hadj Taieb, Thomas Shafee, Tiago Lubiana, Dariusz Jemielniak, Mohamed Ben Aouicha, … Daniel Mietchen. (2020). Representing COVID-19 information in collaborative knowledge graphs: a study of Wikidata.

    Image
    Image
    Image
    1 reply 2 retweets 8 likes
    Show this thread
    Image
    Show this thread
    Undo
  11. Sep 15

    "Collective response to the media coverage of COVID-19 pandemic on and " (Perra et al, JMIR 2020) arxiv:

    Image
    , , and 7 others
    10 retweets 22 likes
    Undo
  12. Sep 14

    Looking for Wikipedia research publications about a particular topic? Try searching the archives of our monthly newsletter (covering over 1000 papers going back to 2011)

    Image
    5 retweets 16 likes
    Undo
  13. Sep 14

    "The Wikipedia Diversity Observatory". A Project to Identify, Quantify and Bridge Content Gaps in ( and , 2020) Paper: Observatory:

    Image
    , , and 2 others
    7 retweets 19 likes
    Undo
  14. Retweeted
    Sep 13

    Wikipedia also lowers financial corruption by reducing insider and manager advantages in the stock market. This paper finds that managers have a harder time holding back bad news for companies with good Wikipedia pages. Having a aggregated source of information eliminates secrets

    Image
    14 retweets 29 likes
    Undo
  15. Sep 11

    "Language Models and Word Sense Disambiguation: An Overview and Analysis" (Loureiro et al, 2020)

    Image
    , , and 4 others
    8 retweets 16 likes
    Undo
  16. Retweeted
    Sep 11

    Continuing theme of great data dump names, also provides "Inspector Javert's Xref Database" using PyOBO, cross-referenced identifiers for Chebi, , PDB, Kegg, HMDB, EFO, GO, HGNC, MESH, Uberon, UMLS, and more 3/

    3 replies 1 retweet 2 likes
    Show this thread
    Image
    Show this thread
    Undo
  17. Retweeted
    Sep 11

    Tough Tables: Carefully Benchmarking Semantic Table Annotators Designed to evaluate table annotation approaches src: SemTab2019 Geodatos by 5/

    1 reply 2 retweets 4 likes
    Show this thread
    Image
    Show this thread
    Undo
  18. Sep 10

    "VideoCutTool - Online Video Editor Tool for Wikimedia Commons" (Vasanth et al, ICCSP 2020) paper: tool: Commons page: GSOC report:

    Image
    , , and
    5 retweets 14 likes
    Undo
  19. Sep 9

    A benchmark for knowledge-intensive language tasks (KILT), all grounded in the same snapshot of .

    Image
    I'm excited to present KILT, a benchmark for Knowledge Intensive Language Tasks. 11 datasets, 5 task families, a unified knowledge source and interface -> . Collaboration b/w 🚀 [1/6]
    Show this thread
    8 retweets 21 likes
    Undo
  20. Retweeted
    Sep 9
    Replying to

    If you're interested in the coverage of , please check out this preprint "Using logical constraints to validate information in collaborative knowledge graphs: a study of COVID-19 on Wikidata"

    4 retweets 4 likes
    Undo
  21. Retweeted
    Sep 8

    The upside of participation trophies: small tokens of achievement can result in big changes in behavior. Giving participation badges to new editors on Wikipedia increases long-term retention by 20%. And it happens even though accounts are pseudonymous & not tied to real names!

    Image
    Image
    13 retweets 40 likes
    Undo

@WikiResearch hasn't Tweeted yet.

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

    You may also like

    ·