addressalign-toparrow-leftarrow-rightbackbellblockcalendarcameraccwcheckchevron-downchevron-leftchevron-rightchevron-small-downchevron-small-leftchevron-small-rightchevron-small-upchevron-upcircle-with-checkcircle-with-crosscircle-with-pluscrossdots-three-verticaleditemptyheartexporteye-with-lineeyefacebookfolderfullheartglobegmailgooglegroupshelp-with-circleimageimagesinstagramFill 1linklocation-pinm-swarmSearchmailmessagesminusmoremuplabelShape 3 + Rectangle 1ShapeoutlookpersonJoin Group on CardStartprice-ribbonShapeShapeShapeShapeImported LayersImported LayersImported Layersshieldstartickettrashtriangle-downtriangle-uptwitteruserwarningyahoo

Similarity and Mutual Information Papers Discussion

We're going to try doing a discussion around two papers. They are not overly technical, but provide a good basis for thinking about a lot of different applications.

 

Papers

Dekang Lin's An Information-Theoretic Definition of Similarity

Peter Turney's Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews

The two papers build off each other. The first discusses mutual information as a method of indicating similarity, and the second applies mutual information for determining the sentiment of phrases.

 

Discussion

Phillip and Lee have agreed to lead the discussion for the two papers.

Reading groups clearly work better when everyone actually reads the paper, so we're going to try a little social pressure. If you are coming, please sign up to be notified of comments. When you finish reading one of the papers please add a comment that you read it.

Its not bragging, its encouraging. 

 

Anticipated Schedule
6:00PM: Network and Socialize
6:30PM: Intros/Announcements
6:40PM: Discussion
7:30PM: More Socializing (presumably followed by dinner/drinks, location TBD)

 

Join or login to comment.

  • Greg B.

    Thanks everyone for coming. I've got to say the reading group went a lot better than I was expecting it to, and I found it really enjoyable. Thanks everyone for engaging so deeply in the discussion. If anyone has suggestions for future papers to read please suggest them. Even better if you want to present them.

    June 18, 2013

  • A former member
    A former member

    I also really like the discussion and as someone who is new to NLP I appreciated having both a theoretical and applied paper side by side!

    June 18, 2013

  • Anna

    I found this very useful. Good papers, great discussion during & post meeting!

    June 18, 2013

  • Jay S.

    Good selection of papers to discuss and some interesting discussion around what the authors were contributing.

    June 18, 2013

  • A former member
    A former member

    I've read the papers. Looking forward to the discussion!

    1 · June 17, 2013

  • amber

    By the way, I recommend this paper by Hirst and Budanitsky that discusses several measures of semantic similarity, including Dekang Lin's: http://ftp.cs.toronto.edu/pub/gh/Budanitsky+Hirst-2006.pdf

    3 · May 17, 2013

    • Greg B.

      Thanks for the link to that paper. Didn't read it all yet, but even just the introductory discussion of similarity vs relatedness is really interesting. And good to see Lin's method independently compared to others.

      1 · June 16, 2013

  • Greg B.

    Way behind on reading, but finished up Lin's paper yesterday and just starting Turney's paper now.

    June 16, 2013

  • Bruce C.

    Kinda new to NLP, but I am very interested in it! I will definitely read the papers this week!

    June 10, 2013

  • amber

    My flight doesn't get back from Atlanta until 7 pm that evening. Any chance we could postpone?

    May 17, 2013

    • Greg B.

      Unfortunately probably not. We had a tough time finding a date that worked for both Phillip, Lee and I. June is a somewhat crazy month.

      May 22, 2013

    • amber

      Too bad! Hopefully I'll see you next time.

      May 22, 2013

  • Cynthia K.

    i'm so bummed. I would love to attend. Out of Town that week.

    May 17, 2013

19 went

Sign up

Meetup members, Log in

By clicking "Sign up" or "Sign up using Facebook", you confirm that you accept our Terms of Service & Privacy Policy