BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Processing Geographic Language - Dr. Inderjeet Mani
DTSTART:20090522T110000Z
DTEND:20090522T120000Z
UID:TALK18301@talks.cam.ac.uk
CONTACT:Johanna Geiss
DESCRIPTION:Humans are able to communicate geographic information in a\nhi
 ghly concise but vague manner\, posing interesting challenges for\nnatural
  language understanding. In recent years\, information\nextraction systems
  have been developed to ground geographical\nreferences in text in terms o
 f geo-coordinates\, with the tags produced\nby such systems being used by 
 geographical search engines and mapping\ntools. However\, without a standa
 rd for how different types of\ngeographical entities should be tagged\, su
 ch systems are impossible to\nreliably evaluate. I will describe an annota
 tion scheme called\nSpatialML\, that has been used to accurately mark up p
 laces\, their\ngeo-coordinates\, and spatial relationships in a variety of
  text\ncorpora. SpatialML represents spatial relationships among geographi
 cal\nregions in terms of the Region Connection Calculus (RCC)\, and it has
 \nalso been mapped to the Generalized Upper Model (GUM) ontology from\nthe
  University of Bremen. SpatialML is also being used in the\nCross-Language
  Evaluation Forum (CLEF) to assess tools to analyze\ngeographical queries 
 posed to search engines\, and it is currently\nbeing integrated with a tim
 e markup standard (TimeML). Despite these\npositive trends\, I will argue 
 that a far more concerted research\neffort is required to address thefunda
 mental challenges of geographic\nlanguage.\n\nDr. Inderjeet Mani is a Visi
 ting Fellow at Cambridge. He has been\na Senior Principal Scientist at MIT
 RE (in Boston)\, a Research Scholar\nat Brandeis University\, a Research A
 ffiliate at MIT\, and a (tenured)\nAssociate Professor at Georgetown Unive
 rsity. His research areas in\nnatural language processing include automati
 c summarization\, temporal\nand spatial information extraction\, and narra
 tive understanding.  More\ninformation can be found at www.cs.brandeis.edu
 /~im5.
LOCATION:SW01\, Computer Laboratory
END:VEVENT
END:VCALENDAR