Submissions/Lucid Links

From Wikimania 2013 • Hong Kong
Jump to navigation Jump to search
Xmark01.svg

After careful consideration, the programme committee has decided not to accept the below submission at this time. Thank you to the author(s) for participating in the Wikimania 2013 programme submission, we hope to still see you at Wikimania this August.

Submission no.
4009
Subject no.
C1
Title of the submission
Lucid Links
Type of submission
presentation
Author of the submission
Oren Bochman
Country of origin
Israel
Affiliation
E-mail
orenbochman@gmail.com
Username
OrenBochman
Personal homepage or blog
Abstract

Part of the growth of wikis is their development of scale-free link graphs (due to editors' bias to link to common memes). Users on a wiki browse by following many links which can be modeled as a random walk over the link graph. However, in terms of information efficiency, stepping over a scale-free graph is neither effective nor conducive to a pleasant user experience. Lucid Links rank existing and potential links in an article. It offers a Manual of Style-compliant recommendation for optimal wikification, and can be used to improve the information efficiency of a wiki.

Detailed proposal

Part of the growth of Wikis is their development of scale-free link graphs (due to editors' bias to link to common memes). Users on a wiki browse by following many links which can be modeled as a random walk over the link graph. However, in terms of information efficiency, stepping over a scale-free graph is neither effective nor conducive to a pleasant user experience. Lucid Links rank existing and potential links in an article. It offers a Manual of Style-compliant recommendation for optimal wikification, and can be used to improve the information efficiency of a wiki.

Lucid Links combines a search engine, a recommendation engine and a dynamic programming algorithm which can provide an optimal solution to the question of which terms (and in which sentences) should be wikified. It is a tool that analyzes Wikipedia dumps and extracts both full text and internal linking structure of each article.

This tool helps to perform optimal wikification, using Shannon's information theory, graph theory, and a very simple idea called "regularization"

  1. Removes over-wikification per the Manual of Style.
  2. Picks the best links per sentence, considering all the sentences in an article and all possible pages.
  3. Makes a best effort at disambiguation links via a semantic topic map.
  4. The method called regularization would transform the "shape of the wiki" from a scale-free network characterized by a few hubs (nodes like USA) to a more regular graph structure in which wikilinks are primarily directed to closely related articles.
  5. A gadget may be installed which can dynamically wikify all potential links according to their global utility.


Track
  • Analysis and Public Engagement
Length of presentation/talk
25 Minutes
Language of presentation/talk

English

Will you attend Wikimania if your submission is not accepted?

Yes, provided I can secure a suitable scholarship.

Slides or further information (optional)
Special requests


Interested attendees

If you are interested in attending this session, please sign with your username below. This will help reviewers to decide which sessions are of high interest. Sign with four tildes. (~~~~).

  1. അമിർ എ. അഹരൊനി (talk) 20:31, 5 March 2013 (UTC)
  2. Wow, is this possible? Blue Rasberry (talk) 20:35, 1 April 2013 (UTC)
  3. Pgallert (talk) 07:19, 4 April 2013 (UTC)
  4. Slashme (talk) 17:50, 7 April 2013 (UTC)
  5. Daniel Mietchen (talk) 23:08, 22 April 2013 (UTC)
  6. Ijon (talk) 00:21, 16 May 2013 (UTC)