(no subject)

David Powers (powers@ist.flinders.edu.au)
Tue, 14 Oct 1997 13:16:09 +0930

Paradigms and Grounding in Language Learning

Wednesday 21st and Thursday 22nd January
Flinders University of South Australia, Adelaide
(part of the Australian NLP Fortnight)

This workshop is focussed on alternatives to the statistical
and learning methodologies based on tagged or bracketed corpora.
Such learning has the advantage that fast training can be performed
and highly effective systems can be developed for the type of
text on which it has trained and the grammatical formalism with
which the corpus was prepared. On the other hand, it has several
disadvantages: it assumes the very formalism which it is trying
to learn, it depends on statistics to choose between ambiguous
parses, it offers no opportunity to learn a grounded semantics.

This workshop is interested in techniques which automatically
propose segmentations, classifications and parses based on plain
text or speech corpora, whether syntactic, semantic, phonological or
morphological. We are particulary interested in language learning in
embedded environments where syntactic, semantic and pragmatic
information can be learned in a way comparable with the way in
which children learn language. We are also interested in theoretical
issues relating to the learning situation or paradigm, and approaches
which adopt an interdisciplinary approach using models from
in psycholinguistics or cognitive linguistics or cognitive science.

Submissions should be in the ANLPF conference submission format and
may be two to four pages in length. All submissions will be reviewed.
Longer papers will be considered at the discretion of the organizers.
Papers should be submitted in final format in PostScript

Topics may include but are not limited to:

Unsupervised paradigms language learning
Algorithms for unsupervised language learning
Statistical methods for unsupervised language learning
Self-organizing neural nets for language learning
Segmentation and Classification models
Grounded models of language and ontology
Learning or evolution/emergence of robot communication
Learning of taxonomies and semantic networks

Deadline for papers:

14th November 1997

Acceptances will be notified by:

28th Novebmer 1997

Organizer:

David Powers
Loebner Prize Committee
Flinders University

Address for correspondence and postscript submissions:

conll@ai.cs.flinders.edu.au

For further details of the Australian Natural Language Processing
Fortnight and paper submission format see:

http://www.cs.flinders.edu.au/research/AI/ANLPF/

-- 
--	powers@acm.org	 http://www.cs.flinders.edu.au/people/DMWPowers.html
 Associate Professor David M. W. Powers		David.Powers@flinders.edu.au
ACM SIGART Editor, ACL SIGNLL ImPastPresident   Facsimile:   +61-8-8201-3626
 Director, AI Lab, Dept of Computer Science	UniOffice:   +61-8-8201-3663
 The Flinders University of South Australia	Secretary:   +61-8-8201-2662
GPO Box 2100, Adelaide	South Australia 5001	HomePhone:   +61-8-8357-4220