Abstracts

Gheorghe Tecuci
Center for Artificial Intelligence, Department of Computer Science, George Mason University, Farfax, VA 22030, U.S.A; tecuci@cs.gmu.edu and
Center for Machine Learning, Natural Language Processing and Conceptual Modelling, Romanian Academy, Bucharest, Romania

N/A(pp. 327-330)

Keywords: case-based reasoning, machine learning, multi strategy approach, single-strategy learning

Diana Gordon
Naval Research Laboratory, Code 5514, Washington D.C. 20375; gordon@aic. nrl.navy.mil
Devika Subramanian
Department of Computer Science, Cornell University, Ithica, NY 14853; devika@cs.cornell.edu

The problem of designing refining task-level strategies in an embedded multiagent setting is an important unsolved question. To address this problem, we have developed a multistrategy system that combines two learning methods: operationalization of high-level advice provided by a human and incremental, refinement by a, genetic algorithm. The 'first method generates seed rules for finer-grained refinements by the genetic algorithm. Our multistrategy learning system is evaluated on two complex simulated domains as well as with a Nomad 200 robot.(pp. 331-346)

Keywords: multistrategy learning, advice taking, compilation, operationalization, genetic algorithms

Ashwin Ram
College of Computing, Georgia Institute of Technology, Atlanta, Georgia 30332- 0280, U.S.A.
Juan Carlos
Santamaria College of Computing, Georgia Institute of Technology, Atlanta, Georgia 30332-0280, U.S.A.

This paper presents a self-improving reactive control system for autonomous robotic navigation. The navigation module uses a schema-based reactive control system to perform the navigation task. The learning module combines case-based reasoning and reinforcement learning to continuously tune the navigation system through experience. The case-based reasoning component perceives and characterizes the system's environment, retrieves an appropriate case, and uses the recommendations of the case to tune the parameters of the reactive control system. The reinforcement learning component refines the content of the cases based on the current experience. Together, the learning components perform on-line adaptation, resulting in improved performance as the reactive control system tunes itself to the environment, as well as on-line case learning, resulting in an improved library of cases that capture environmental regularities necessary to perform on-line adaptation. The system is extensively evaluated through simulation studies using several performance metrics and system configurations.(pp. 347-369)

Keywords: Robot navigation, reactive control, case-based reasoning, reinforcement learning, adaptive control

Gerhard Widmer
Department of Medical Cybernetics and Artificial Intelligence, University of Vienna, and Austrian Reseaxch Institute for Artificial Intelligence, Schottengasse 3, A- 1010 Vienna, Austria

The paper presents a general learning method that integrates knowledge-based symbolic learning with instance-based numeric learning. This combination is motivated by a class of learning problems where the task is to predict a numeric target variable, and where a qualitative domain theory is available. The method has been implemented in a learning program named IBL-SMART. Its symbolic learning component is a multiinstance plausible explanation algorithm that can use the qualitative domain knowledge to guide its search and the numeric component pez-forms instance-based prediction of the numeric target variables. The system has been applied to a complex problem from the domain of tonal music. The application domain is briefly described, and some experiments are presented to illustrate the power of the learning method.(pp. 371-385)

Keywords: artificial intelligence, multistrategy learning, instance-based learning, knowledge-based learning, qualitative knowledge, music

Paul T. Baffes
Department of Computer Sciences, University of Texas, Austin, Texas 78712-1188 USA
Raymond J. Mooney
Department of Computer Sciences, University of Texas, Austin, Texas 78712-1188 USA

In recent years, machines learning research has started addressing a problem known as theory refinement. The goal of a theory refinement learner is to modify an incomplete or incorrect rule base, representing a domain theory, to make it consistent with a set of input training examples. This paper presents a major revision of the Either propositional theory refinement system. Two issues are discussed. First, we show how run time efficiency can be greatly improved by changing form a exhaustive scheme for computing repairs to an iterative greedy method. Second, we show how to extend Either to refine M - of - N rules the resulting algorithm ,NEITHER( New EITHER ) is more than an order magnitude and produces significantly more accurate results with theories that fit the M of N format. To demonstrate the advantages of Neither, we present experimental results form two real-world domains.(pp. 387-397)

Keywords: Artificial intelligence, multistrategy learning, theory refinement

Michael R.Hieb
Center for Artificial Intelligence, George Mason University, Fairfax, VA; hieb@gmu.edu
Ryszard S. Michalski
Center for Artificial Intelligence, George Mason University, Fairfax, VA; michalski@gmu.edu

Research on multistrategy task-adaptive learning aims at integrating all basic inferential learning strategies---learning by deduction, induction and analogy. The implementation of such a learning system requires a knowledge representation that facilitates performing a multitype inference in a seamlessly integrated fashion. This paper presents an approach to implementing such multitype inference based on a novel knowledge representation, called Dynamic Interlaced Hierarchies (DIH). DIH integrates ideas from our research on cognitive modeling of human plausible reasoning, the Inferential Theory of Learning, and knowledge visualization. In DIH, knowledge is partitioned into a 'static' part that represents a relatively stable knowledge, and a 'dynamic' part that represents knowledge that changes relatively frequently. The static part is organized into type, part, or precedence hierarchies, while the dynamic part consists of traces that link nodes of different hierarchies. By modifying traces in different ways, the system can perform different knowledge transmutations ( patterns of inference), such as generalization, abstraction, similization, and their opposites, specialization, concretion, and dissimilization, respectively.(pp. 399-412)

Keywords: Multistrategy learning, inferential theory of learning, knowledge transmutation, generalization, abstraction, similization