DEEP THOUGHT: Hybrid and Shallow Methods for Knowledge-Intensive
Information Extraction
A postdoctoral research position will shortly be available at the
University of Sussex on the DEEP THOUGHT project, funded by the EU
Framework 5 IST Programme. The position will be available from January
2003, for 21 months.
The project is concerned with integrating shallow and deep approaches to
the automatic analysis of natural language. The aim is to preserve the
advantages of shallow processing (coverage, efficiency) while adding
more accuracy and depth in a controlled fashion at places where an
application has a real need for it. The project will apply the new
technology in three types of knowledge-intensive application:
Information extraction for business intelligence
Email response management for customer relationship management
Creativity support for document production and collective brainstorming
The project consortium is:
Saarland University, Department of Computational Linguistics
Saarbruecken, Germany <http://www.coli.uni-sb.de/>
University of Sussex, Cognitive and Computing Sciences
Brighton, UK <http://www.cogs.susx.ac.uk/lab/nlp/>
University of Cambridge, Computer Laboratory
Cambridge, UK <http://www.cl.cam.ac.uk/Research/NL/>
Norges teknisknaturvitenskapelige universitet, Department of Linguistics
Trondheim, Norway <http://www.ling.hf.ntnu.no/>
CELI s.r.l.
Turin, Italy <http://www.celi.it/english/>
XtraMind Technologies GmbH
Saarbruecken, Germany <http://www.xtramind.com/english/html/home.php>
At Sussex, work will involve:
refining proposals for an underspecifiable semantic representation
(robust minimal recursion semantics, RMRS), and modifying existing
shallow analysis tools for English to produce this representation
investigating approaches for integrating shallow RMRS analysis with
deeper HPSG-based analysis which also produces (R)MRS
using shallow parsing of domain-specific text to automatically acquire
subcategorisation lexica
developing techniques for tactical generation of fluent text from
RMRS within an existing HPSG processing system
The position requires a high level of competence in programming, ideally
in Common Lisp and C (Perl would also be useful), and experience of
statistical or machine learning approaches to natural language processing.
The salary scale is in the range GBP 18265-25451, depending on age and
experience. To apply for the position, please email your CV (plain text,
postscript or PDF) and the names and addresses of three referees to John
Carroll (johnca@cogs.susx.ac.uk).
This archive was generated by hypermail 2b29 : Mon Oct 14 2002 - 11:46:05 MET DST