The workshop aims at bringing together research in two complementary fields of semantic analysis that are still too far apart. In order to achieve both a broad and a deep understanding of any given text document, a system needs both advanced acquisition of corpus specific lexical semantic knowledge and powerful inference mechanisms that utilize that knowledge in discourse analysis.
Given the still relatively limited results within both areas there has been little impetus to combine them. Corpus-based extraction of lexical semantic knowledge has only recently become a more feasible task, because of the growing availibility of on-line text documents; robust corpus processing technologies, such as broad coverage part-of-speech tagging and shallow parsing; and readily available statistical methods. The various approaches to discourse analysis, originating in such diverse fields as formal semantics, psychology and AI, are in the process of converging into a unified approach to the analysis and representation of the cohesive structure of natural language documents.
The intersection between these two fields lies in the application of lexical semantic knowledge to such problems in discourse analysis as anaphora resolution and discourse segmentation. In fact, the benefit will be mutual, because knowledge of discourse structure is helpful to lexical knowledge extraction as well.
In summary, large scale domain specific lexical semantic knowledge acquisition can assist in analyzing discourse structures, which in turn can assist in acquiring even more accurate lexical semantic representations for the relevant terms in the domain.
The workshop will consist of five sessions, with two or three 20+10-minute presentations in each session. Additionally, a number of invited talks will be given by members of the program committee, while also some time will be set aside for general discussion of the topic of the workshop.
- Bran Boguraev
- Ann Copestake
- Daniel Kayser
- Alex Lascarides
- Manfred Pinkal
- Massimo Poesio
- James Pustejovsky
It should be stressed that we especially encourage those contributions that address aspects of the integration of all three areas of research mentioned above (lexical semantics, discourse analysis and corpus-based approaches to both of these).
The accepted papers will be made available in a summer school reader. Publication in an edited volume is under discussion.
All researchers in the area, but especially Ph.D. students and young researchers, are encouraged to submit an extended abstract (4 to 5 pages), preferably by email and in postscript.
Submissions should be sent before February 15, 1998 to one of the following two organizers:
Johan Bos ( bos@coli.uni-sb.de )
University of the Saarland Dept. of Computational Linguistics Postfach 15 11 50 D-66041 Saarbruecken, Germany |
Paul Buitelaar ( paulb@cs.brandeis.edu )
Brandeis University Computer Science Department 415 South Street Waltham MA 02254, USA |
March 2, 1998 : Deadline for submissions April 15, 1998 : Notification of acceptance May 15, 1998 : Deadline for final copy August 17, 1998 : Start of workshop