NC STATE UNIVERSITY
		       DEPT OF COMPUTER SCIENCE
			       SEMINAR

     Using Plan Reasoning in the Generation of Plan Descriptions

			    Michael Young
				CMU
			 (Faculty candidate)


			   in 402A Withers
			      at 3:30pm
			  on Jan 30th, 1998


			       Abstract

Intelligent systems are often called upon to form plans that specify
their own or other agents' activities.  For these systems, the ability
to describe plans to people in natural ways is an essential aspect of
their interface.  There is a mismatch, however, in the level of detail
present in the simplest of plans and the detail found in people's
descriptions of them.  As a result, computer systems responsible for
automatically generating descriptions of plans must pay careful
attention to the interaction between the quality of a description and
the quantity of information it contains.

In the first part of the talk, I'll present Clia, the Cooperative Plan
Identification Architecture --- a computational model that generates
concise textual descriptions of plans.  Clia generates plan
descriptions by heuristic search, using a model of the user's plan
reasoning process, her plan-related preferences and her reasoning
resource bounds to produce descriptions that are at once concise and
effective.  This approach defines a computational model in which
computer systems and human users are implicit collaborators in their
communication about a plan.  A user interprets a concise plan
description by filling in the missing detail using plan reasoning.  A
cooperative system selects the content of a plan description based on
an expectation that the user is able to complete the description in
much the same way that a planning system completes a partial plan.

The Clia architecture has been empirically evaluated in an experiment
which I'll describe in the second part of the talk.  In this
experiment, human subjects were given written instructions produced by
Clia and by alternative approaches; the subjects were then asked to
carry out the instructions within a virtual environment.  The results
of this experiment clearly showed that subjects following instructions
produced by Clia performed their tasks with fewer execution errors and
achieved a higher percentage of their tasks' goals than did subjects
following instructions produced by the alternative methods.


Host: James Lester