Thedevelopmentofmultistrategylearningsystemsrequiresaclearunderstandingoftherolesandthe applicabilityconditionsofdifferentlearningstrategies.Tothisend,thischapterintroducesthe InferentialTheoryofLearning thatprovidesaconceptualframeworkforexplaininglogicalcapabilities oflearningstrategies,i.e.,their competence.Viewinglearningasaprocessofmodifyingthelearner's knowledgebyexploringthelearner'sexperience,thetheorypostulatesthatanysuchprocesscanbe describedasasearchina knowledgespace, which involvesthelearner'sexperience,piorknowledgeand the learninggoal .Thesearchoperatorsareinstantiationsof knowledgetransmutations, whichare genericpatternsofknowledgechange.Transmutationsmayemployanybasictypeofinference --- deduction,inductionoranalogy.Severalfundamentalknowledg etransmutationsaredescribedinanovel andgeneralway,suchasgeneralization,abstraction,explanationandsimilization,andtheircounterparts, specialization,concretion,predictionanddissimilization,respectively.Generalizationenlargesthe referenceset ofadescription(thesetofentitiesthatarebeingdescribed).Abstractionreducesthe amountofthedetailaboutthereferenceset.Explanationgeneratespremisesthatexplain(orimply)the givenpropertiesofthereferenceset.Similization transfersknowledgefromonereferencesettoasimilar referenceset.Usingconceptsofthetheory,a multistrategytask -adaptivelearning(MTL)methodology isoutlined,andillustratedbyanexample.MTLdynamicallyadaptsstrategiestothe learningtask , definedbytheinputinformation,learner'sbackgroundknowledge,andthelearninggoal. Thegoalof MTLresearchisto synergisticallyintegrateawiderangeofinferentiallearningstrategies,suchas empiricalgeneralization,constructiveinduction, deductivegeneralization,explanation,prediction, abstraction,andsimilization. Keywords: learningtheory,inferencetheory,multi...

The software and knowledge engineering literature defines maintenance strategies for seven main types of knowledge: words; sentences; behavioural knowledge; and meta-knowledge. Meta-knowledge divides into problem solving methods; quality knowledge; fix knowledge; social knowl- 5 edge; and processing activities. There are five main ways in which these seven knowledge types are processed: acquire; operationalise; fault; fix; and preserve. We review systems that contribute to these 7 5 = 35 types of knowledge maintenance. 1 Introduction 10 A general trend in the twentieth century is an increasing level of doubt about the things we speak or write or try to enter into programs. Popper argues that all knowledge is an hypothesis since nothing can ever be ultimately proved; Submitted to the Knowledge Engineering Review page 2 of 73 our currently believed ideas are merely those that have survive active attempts to refute them [89]. Knowledge representation theorists stress that KBs are...

The paper presents an approach to the automation of knowledge acquisition for expert systems. The approach is based on several general principles emerging from the field of machine learning: expert system building as a three phase process, understanding-based knowledge extension, knowledge acquisition through multistrategy learning, consistency-driven concept formation and refinement, closed-loop learning. and synergistic cooperation between the human expert and the learning system. In this approach, an expert system is built by a human expert and a learning system. The human expert defmes the framework for the expert system and provides an incomplete and partially incorrect knowledge base. The learning system incrementally extends, updates, and improves the knowledge base through learning from the human expert. This approach is illustrated by the learning system shell NeoDISCIPLE.

The purpose of this research is to extend the theory of uncertain reasoning over time through integrated, multi-strategy learning. Its focus is on decomposable, concept learning problems for classification of spatiotemporal sequences. Systematic methods of task decomposition using attribute-driven methods, especially attribute partitioning, are investigated. This leads to a novel and important type of unsupervised learning in which the feature construction (or extraction) step is modified to account for multiple sources of data and to systematically search for embedded temporal patterns. This modified technique is combined with traditional cluster definition methods to provide an effective mechanism for decomposition of time series learning problems. The decomposition process interacts with model selection from a collection of probabilistic models such as temporal artificial neural networks and temporal Bayesian networks. Models are chosen using a new quantitative (metric-based) approach that estimates expected performance of a learning architecture, algorithm, and mixture model on a newly defined subproblem. By mapping subproblems to customized configurations of probabilistic networks for time series learning, a hierarchical, supervised learning system with enhanced generalization quality can be automatically built. The system can improve data fusion

This memory can take several forms based on the time scales at which such modifications are allowed. Some symbol structures might have the property of determining choice and the order of application of transformations to be applied on other symbol structures. These are essentially the programs. Programs when executed --- typically through the conventional process of compilation and interpretation and eventually --- when they operate on symbols that are linked through grounding to particular effectors --- produce behavior. Working memory holds symbol structures as they are being processed. Long-term memory, generally speaking, is the repository of programs and can be changed by addition, deletion, or modification of symbol structures that it holds. Such a system can compute any Turing-computable function provided it has sufficiently large memory and its primitive set of transformations are adequate for the composition of arbitrarily symbol structures (programs) and the interpreter is capable of interpreting any possible symbol structure. This also means that any particular set of symbolic processes can be carried out by an NANN --- provided it has potentially infinite memory, or finds a way to use its transducers and effectors to use the external physical environment to serve as its memory). 14 Chapter 12 Knowledge in SAI systems is typically embedded in complex symbol structures such as lists (Norvig, 1992), logical databases (Genesereth and Nilsson, 1987), semantic networks (Quillian, 1968), frames (Minsky, 1975), schemas (Arbib, 1972; 1994), and manipulated by (often serial) procedures or inferences (e.g., list processing, application of production rules (Waterman, 1985), or execution of logic programs (Kowalski, 1977) carried out by a central processor that accesse...

1 An understanding of learning -- the process by which a learner acquires and refines a broad range of knowledge and skills -- is central to the enterprise of building truly adaptive, flexible, robust, and creative intelligent systems. Significant theoretical and empirical contributions to the characterization of learning in computational terms have emerged from research in a number of disparate research paradigms. The limitations of individual paradigms and of particular classes of techniques within each paradigm are beginning to be recognized. Converging lines of evidence from multiple sources, both theoretical as well as empirical, suggest that artificial intelligence systems, in order to be able to deal with complex tasks such as recognizing and describing 3-dimensional objects, or communicating in natural language, must be able to effectively utilize a range of learning algorithms operating with an adequate repertoire of representational structures. This paper draws on a broad ran...

This paper outlines basic assumptions and a theoretical basis for multistrategy task.adaptive learning (MTL) methodology, which aims at ultimately integrating a spectrum of learning strategies, such as empirical teaming, constructive induction, abduction, analytic learning. learning by analogy, and reinforcement learning. In MTL, in response to an input, a learner deternines the su'ategy, or a combination of su'ategies, that is mo. st appropriate for the learning task. This detemination is based on the relationship between the input, the leamegs background knowledge and the leamer's task. By means of a simple example we show how an MTL learner can employ, depending on the above relationship, emprical learning, constructive inductive generaiiz. afion, abduction, explanation-based learning and abstraction.

Initial results toward developing a unifying conceptual framework for characterizing diverse learning strategies and paradigms are presented. We outline the inferential theory of learning that aims at understanding competence aspects of learning processes, in contrast to computational theory that is concerned with complexity aspects. The theory views learning as a goal-oriented process of creating or modifying knowledge representations. Such a process may involve any type of inference (deduction, analogy or induction) or information transmutation (e.g., reformulation, abstraction or copying). Any type of learning can therefore be characterized in terms of the types of such knowledge transformations that occur in a learning process. Several concepts fundamental to understanding learning are analyzed in a novel way and compared, such as analytic vs. synthetic learning, deduction, induction, abduction, abstraction and generalization. It is shown, for example, that inductive generalization, inductive specialization and abduction can be viewed as various forms of general induction, and that abstraction is a form of constructive deduction. Based on these concepts, a general multicriterion classification of learning processes is proposed. The presented ideas have a special significance for the development of a new generation of learning systems, called multistrategy systems, that integrate diverse learning strategies in a goal-oriented fashion. 1.

Among the central tasks in the development of expert systems is the formulation, debugging and implementation of a knowledge base. The knowledge encoded in the knowledge base is usually supplied by experts. There are, however, many application domains in which knowledge required by an expert system has to be extracted from facts collected in a data base. In view of the large sizes and the complexity of contemporary data bases in different areas, such as agriculture, medicine, business, etc., determining useful knowledge from them is becoming an increasingly difficult problem. This paper describes a multistrategy "intelligent assistant" for knowledge discovery in large data bases, called INLEN. The system integrates a database, a knowledge base, and machine learning capabilities within a uniform user-oriented framework. The latter capabilities are incorporated into the system in the form of knowledge generation operators (KGOs). These operators can generate diverse kinds of knowledge about the properties and mgularities existing in the dam. For example, they can hypothesize general diagnostic rules from specific cases of diagnosis, optimize the rules according to problem-dependent criteria, determine differences and similarities among groups of facts, propose oew variables, create conceptual classifications, determine equations governing numeric variables and the conditions under which the equations apply, derive statistical properties and use them for qualitative evaluations, etc. The initially implemented system, INLEN-1, is described, and its performance is illustrated by an example.

This paper presents a learning-based approach to the automation of knowledge acquisition for expert systems. An expert system is viewed as an explicit mooel of a human expert's competence and perfonnance. We distinguish three phases in the development of such a model. The fIrst one consists of defIning a framework for the mooel, in terms of a knowledge representation formalism and an associated problem solving methoo. The second phase consists of defIning a preliminary mooel that describes the basic concepts of the expertise domain. The last phase consists of incrementally extending and improving the domain model through learning from the human expert. The paper describes the learning system NeoDISCIPLE which illustrates the usefulness of six principles for automating the knowledge acquisition process: expert system building as a threephase mooeling of human expertise, understanding-based knowledge extension, knowledge acquisition through multistrategy learning, consistency-driven concept fonnation and refinement, closed-loop learning, and cooperation between the human expert and the learning system.