Adaptive Reasoning for Real-World Problems:
A Schema-Based Approach

Roy M. Turner

Department of Computer Science
University of New Hampshire

This book describes a method for building real-world problem solving systems, such as medical diagnostic systems and intelligent controllers for autonomous underwater vehicles (AUVs) and other robots. The approach is different from other work reported in the artificial intelligence (AI) literature in several respects. First, it defines schema-based reasoning, in which (explicitly declared packets of related knowledge) are used to control not only the reasoner's planning, but also all other facets of its behavior. Second, the approach is an kind of reactive reasoning that the author call's adaptive problem solving: the reasoner maintains commitments to future goals to be achieved, but it is able to respond appropriately and quickly to unanticipated events, and it is able to change its focus of attention as the problem-solving situation requires. Third, it is a context-sensitive reasoning method. Every decision it makes relies on the use of contextual knowledge to be appropriate for the current problem-solving situation. Furthermore, context is represented explicitly; by always keeping a current representation of the context ``in mind,'' the reasoner's behavior is automatically sensitive to the context with very little work needed per decision. Fourth, schema-based reasoning is a generalization of case-based reasoning; it extends the usual idea of case-based reasoning to encompass all aspects of the reasoner's behavior, and it extends it to make use of generalized ``cases'' (i.e., schemas) rather than particular cases, thus saving effort needed to transfer knowledge from an old case to a new situation.

Though the work originated in the domain of medical diagnostic problem solving, treating diagnosis as a planning task, it is even more appropriate for controlling autonomous systems. The author is currently extending the approach by creating a robust controller for long-range autonomous underwater vehicles that will be used to carry out ocean science missions.

Some Features of Schema-Based Reasoning

1. Schema-based reasoning is a reactive problem solving approach, but one that always reacts appropriately for the context it is in with minimum effort at the time the reaction is needed. (Chapter 2)

2. Schema-based reasoning is a context-sensitive reasoning mechanism that makes use of schemas representing contextual knowledge (contextual schemas, or c-schemas) to automatically behave appropriately for the situation it is in. (Chapter 3)

3. Unanticipated events are handled such that their meaning, their importance, and the appropriate response is determined in a context-specific manner. (Chapter 6)

4. The reasoner's attention is changed as the situation changes using context-specific information to assure that the reasoner is always working on the most appropriate goal. (Chapter 4)

5. The reasoner uses procedural schemas to achieve its goals; p-schemas are flexible structures whose steps and step ordering can be changed to fit the situation in which they are applied. They are expanded only as much as needed (i.e., a least-commitment approach); by deciding on future actions near the time they will be taken, the reasoner can adapt its actions to the current problem-solving situation at the time when it knows what that situation is. (Chapter 5)

6. All the reasoner's problem-solving knowledge is represented as schemas, which are organized in generalization-specialization hierarchies. By traversing these hierarchies using features of the current situation, the most appropriate schema or set of schemas is brought to bear on the problem at hand. As the problem situation changes, or as later steps in a p-schema need to be expanded in a changed situation, the reasoner automatically selects the most appropriate knowledge it has. (Chapter 3; Chapter 7)

7. Two implementations of schema-based reasoning are discussed: MEDIC, a pulmonology consultant; and Orca, a controller for a long-range autonomous underwater vehicle that will be used on actual ocean science missions. (Chapter 8)

8. The work is evaluated and compared to related work in the field. (Chapter 9)

Some Features of the Book

The book has 61 figures illustrating schemas, processes, and algorithms. Two appendices are included: an annotated script of MEDIC's output while solving a problem, and a glossary of medical terms used. A bibliography of 130 references is included.

About this document ...

This document was generated using the LaTeX2HTML translator Version 0.5.2 (Mon Jan 25 1994) Copyright © 1993, Nikos Drakos, Computer Based Learning Unit, University of Leeds.

The command line arguments were:
latex2html -split 0 abstract.tex.

The translation was initiated by rmt@cdps.umcs.maine.edu on Fri Jul 7 09:41:42 EDT 1995