AIC Seminar Series
Can Linked Data Help Hypertext Search?
|Harry Halpin||University of Edinburgh||[Home Page]|
Notice: Hosted by Vinay Chaudhri. Webex: voice by telecon: 1-888-355-1249, passcode 749045. WebEx: https://sri.webexone.com/webservice/wxr.aspx?_command=join&MK=483042483
Date: 2010-03-26 at 15:00
Location: EJ228 (SRI E building) (Directions)
Linked Data represents the explosion of structured data
given in the Semantic Web format RDF, released on to the real-existing
hypertext Web. This data can be searched and indexed just like
ordinary data using special purpose "Linked Data" search engines.
Using a real query log from Microsoft Live and named entity and
concept recognition, we show what kinds of queries can and cant be
answered by Linked Data, as well as how Linked Data demonstrates what
features of RDF and OWL are actually used "in the wild". Then we
present the first Cranfield-style evaluation of "semantic search"
(the search for Linked Data), and then show how relevance feedback
(using both vector-space and language modelling approaches) from
hypertext search can improve semantic search, and - more interesting -
vice versa. Lastly, we present upcoming work of us in co-operation
with Yahoo! Research on using Mechanical Turk to "crowd-source" a
large-scale evaluation of semantic search engines.
Im a researcher and postgraduate student of Henry S. Thompson and Andy Clark at the University of Edinburgh. Im interested in the intersection of philosophy and the Web. In particular, what exactly are the secrets to the success of the Web and what lessons does this hold for computing in general, and especially artificial intelligence? In my current work I am analyzing both the underlying architecture of the Web using notions from information theory, type theory, the philosophy of computation. As a side interest, Im also interested in how we can use machine-learning and natural language processing to make the massive amount of text on the Web easier to use, in particular through automatic ontology creation. This application comes from my interest in narrative analysis, which I did my MSc. thesis. I can usually be found in Edinburgh, but have been known to make guest appearances in New York City and the forests of North Carolina.
Please arrive at least 10 minutes early as you will need to sign in by
following instructions by the lobby phone at Building E. SRI is located
at 333 Ravenswood Avenue in Menlo Park. Visitors may park in the parking
lots off Fourth Street. Detailed directions to SRI, as well as maps, are
available from the Visiting AIC web page.
There are two entrances to SRI International located on Ravenswood Ave.
Please check the Builing E entrance signage.
©2014 SRI International 333 Ravenswood Avenue, Menlo Park, CA 94025-3493