| Diagrammatic
Reasoning at the University of New South Wales Norman Y. Foo
The Knowledge Systems Group of the
Department of Artificial Intelligence, School of Computer Science and Engineering, The
University of New South Wales in Sydney, Australia, has started a project on diagrammatic
reasoning. The basis of this work is outlined in one of the sections of a 1997 IJCAI paper
[2] that can be obtained by ftp from ftp://ftp.cse.unsw.
edu.au/pub/users/ksg/Conference/ijcai97_strips.ps.z. The basic idea is to use a form of
the Gaifman distance [3] applied to action pre-conditions and post-conditions to determine
the extent of local structures relevant to diagrams that can be used to reason about the
effects of actions. Such reasoning can be confined to these structures instead of the
larger domain. Work is progressing on testing these concepts in the context of using
diagrams to capture semantics for imperative program constructs applied to data
structures, and to combining diagrammatic information with logic within an object-oriented
framework. Eventually, we hope to use this experience to shed light on why diagrams
succeed when they do, and under what circumstances they may fail. We see this work as
contributing to the fulfilment of the aims first enunciated by Barwise, Etchemendy and
their colleagues [1] which they call "heterogeneous reasoning". This research is
supported by the Australian Research Council, but the challenge of commercial applications
is highly attractive. We invite interested colleagues to correspond with us. Contact:
Norman Foo (norman@cse.unsw.edu.au) or Maurice Pagnucco (morri@cse.unsw.edu.au).
References
[1] G. Allween and J. Barwise
(editors). Working papers on diagrams and logic. Indiana University Logic Group Preprint
Series, No. IULG-93-24, Indiana University, 1993.
[2] N. Foo, A. Nayak, M. Pagnucco, P.
Peppas, and Y. Zhang. Action localness, genericity and constraints in {STRIPS}. Proceedings
of the Fifteenth International Joint Conference on Artificial Intelligence, M. E. [3]
Pollack (Ed), pp. 549--554, Nagoya, Japan, 1997.
[3] H. Gaifman. On local and nonlocal
properties. Logic Colloquium '81, J. Stern (ed), pp. 105--135, North Holland, 1982
Norman Y. Foo
Knowledge Systems Group
Department of Artificial Intelligence
School of Computer Science &
Engineering
University of New South Wales
NSW, 2052, Australia.
Email: norman@cse.unsw.edu.au |