Jump to navigation Jump to search
This is a rejected submission for Wikimania 2012.
- Submission no.
- Title of the submission
- TREC KBA - Mining Content Streams to Recommend Page Updates to Editors
- Type of submission (workshop, tutorial, panel, presentation)
- Author of the submission
- John R. Frank and Daniel A. Roberts
- E-mail address
- jrf at mit dot edu
- Country of origin
- Affiliation, if any (organization, company etc.)
- Personal homepage or blog
- Abstract (at least 300 words to describe your proposal)
- We have organized a new session in NIST's Text Retrieval Conference (TREC) called Knowledge Base Acceleration (KBA). TREC KBA challenges computer science researchers to develop algorithms that mine content streams, such as news and blogs, to recommend edits to knowledge bases (KB), such as Wikipedia. We consider a KB to be "large" if the number of entities described by the KB is larger than the number of humans maintaining the KB. As entities change and evolve in the real world, large KBs often lag behind by months or years. Such large KBs are an increasingly important tool in several industries, including biomedical research, law enforcement, and financial services. TREC KBA aims to develop algorithms for helping KB editors stay abreast of changes to the organizations, people, proteins, and other entities described by their KBs. In this technical presentation, we will give an overview of the TREC KBA data sets and tasks for 2012 and future years. In addition to developing text analytics, we are also working on a wikipedia bot for connecting KBA-type systems to users' talk pages in mediawiki. After presenting the current state of our bot development, we hope to engage the audience in an open discussion about how such algorithms might be most fruitfully employed in the Wikipedia community.
- Technology and Infrastructure:
- Length of presentation/talk
- 25 Minutes
- Will you attend Wikimania if your submission is not accepted?
- Slides or further information (optional)
- Special request as to time of presentations
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. (~~~~).