dlm@research.att.COM (01/22/88)
Generate, Test and Debug: a Paradigm for
Solving Interpretations and Planning Problems
Reid Simmons
Massachusetts Institute of Technology
January 21, 1988, AT&T Bell Labs-Murray Hill 3D-473, 11:00 am
January 22, 1988, AT&T Bell Labs-Holmdel 4C-513, 10:30 am
ABSTRACT
We describe how the Generate, Test and Debug (GTD) paradigm
solves interpretation and planning problems, and why its
combination of associational and causal reasoning
techniques enables it to achieve efficient and robust
performance. The Generator constructs hypotheses using
domain dependent rules. The Tester verifies hypotheses and
supplies the Debugger with causal explanations if the test
fails, and the Debugger uses domain-independent algorithms to
repair hypotheses by analyzing the causal explanations and
models of the domain. The GTD paradigm has been implemented
and tested in the domains of geologic interpretation, the
blocks world, and the Tower of Hanoi problem.
Sponsor: Ron Brachman