Inferential Theory of Learning: Developing Foundations for Multistrategy Learning
Published 1 October 1992Open access
Ryszard S. Michalski
Citations109
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This chapter introduces the Inferential Theory of Learning that provides a conceptual framework for explaining logical capabilities of learning strategies, i.e., their competence, and a multistrategy task-adaptive learning (MTL) methodology, which dynamically adapts strategies to the learning task.
Abstract
The development of multistrategy learning systems requires a clear understanding of the roles and the applicability conditions of different learning strategies. To this end, this chapter introduces the Inferential Theory of Learning that provides a conceptual framework for explaining logical capabilities of learning strategies, i.e., their competence. Viewing learning as a process of modifying the learner’s knowledge by exploring the learner’s experience, the theory postulates that any such process can be described as a search in a knowledge space, which involves the learner’s experience, pior knowledge and the learning goal. The search operators are instantiations of knowledge transmutations, which are generic patterns of knowledge change. Transmutations may employ any basic type of inference— deduction, induction or analogy. Several fundamental knowledge transmutations are described in a novel and general way, such as generalization, abstraction, explanation and similization, and their counterparts, specialization, concretion, prediction and dissimilization, respectively. Generalization enlarges the reference set of a description (the set of entities that are being described). Abstraction reduces the amount of the detail about the reference set. Explanation generates premises that explain (or imply) the given properties of the reference set. Similization transfers knowledge from one reference set to a similar
Keywords
Computer Science
Choice Reviews OnlineGenetic algorithms in search, optimization, and machine learning
49,278 Citations1989
This book brings together the computer techniques, mathematical tools, and research results that will enable both students and practitioners to apply genetic algorithms to problems in many fields.
Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference
16,927 Citations1988Judea Pearl
The author provides a coherent explication of probability as a language for reasoning with partial belief and offers a unifying perspective on other AI approaches to uncertainty, such as the Dempster-Shafer formalism, truth maintenance systems, and nonmonotonic logic.
Philosophy and Phenomenological ResearchObjective Knowledge, an Evolutionary Approach.
3,820 Citations1973Harry Ruja, Karl R. Popper
Artificial IntelligenceThe knowledge level
2,484 Citations1982Allen Newell
A theory of the nature of knowledge is proposed, namely, that there is another computer system level immediately above the symbol (or program) level and knowledge itself is the processing medium at this level and the principle of rationality plays a central role.
Springer series in statisticsMeasures of Association for Cross Classifications
2,260 Citations1979Leo A. Goodman, William Kruskal
Machine LearningExplanation-Based Generalization: A Unifying View
1,175 Citations1986Tom M. Mitchell, Richard M. Keller +1 more
This paper proposes a general, domain-independent mechanism, called EBG, that unifies previous approaches to explanation-based generalization, and is illustrated in the context of several example problems, and used to contrast several existing systems for explanation- based generalization.
Mathematics and Plausible Reasoning
856 Citations1954George Pólya
Computer CompactsA theory and methodology of inductive learning
778 Citations1983Ryszard S. Michalski
Machine LearningChunking in Soar: The Anatomy of a General Learning Mechanism
510 Citations1986John E. Laird, Paul S. Rosenbloom +1 more
The anatomy of chunking in Soar is examined and a new demonstration of its learning capabilities involving the acquisition and use of macro-operators is provided.
Cognitive ScienceThe Logic of Plausible Reasoning: A Core Theory
361 Citations1989Allan Collins, Ryszard S. Michalski
This is one of the first attempts to construct a formal theory that addresses both the semantic and parametric aspects of the kind of everyday reasoning that pervades.
Shift of bias for inductive concept learning
288 Citations1984Paul E. Utgoff
It is shown that search for appropriate bias is itself a major part of the learning task, and that mechanical procedures for conducting a well-directed search for an appropriate bias can be created.
Journal of Logic and ComputationOn the Relationship Between Abduction and Deduction
252 Citations1991Luca Console, Daniele Theseider Dupré +1 more
A meta-level definition of abduction in l d terms of deduction, similar to various definitions proposed in the literature, and an object-leve efinition in which abductive conclusions are expressed as a logical consequence of the obser.
Computational IntelligenceExplanation and prediction: an architecture for default and abductive reasoning
212 Citations1989David Poole
This paper elaborates on the idea that reasoning should be viewed as theory formation where logic tells us the consequences of the authors' assumptions, and proposes an architecture to combine explanation and prediction into one coherent framework.
Machine LearningLearning at the Knowledge Level
202 Citations1986Thomas G. Dietterich
The paper analyzes two classes of learning programs, called symbol level learning and nondeductive knowledge level learning, and speculates on the possibility of developing coherent theories of each.
Cambridge University Press eBooksA HISTORY OF THE INDUCTIVE SCIENCES
192 Citations2010William Whewell
Machine LearningHypothesis-Driven Constructive Induction in AQ17-HCI: A Method and Experiments
187 Citations1994Janusz Wnek, Ryszard S. Michalski
The developed AQ17-HCI system consistently outperformed, in terms of the prediction accuracy on new examples, all systems that it was compared to, including the AQ15 rule learning system, GREEDY3 and GROVE decision-list learning systems, and REDWOOD and FRINGE decision-tree learning systems.
The MIT Press eBooksQuantitative Results Concerning the Utility of Explanation-Based Learning
175 Citations1995Steven Minton
Cognitive ScienceIntegrated Learning: Controlling Explanation
150 Citations1986Michael Lebowitz
A combination of the two methods—applying explanation-based techniques during the course of similiarity-based learning can achieve the power of explanationbased learning without some of the computational problems that can otherwise arise in domains lacking detailed explanatory rules.
Artificial IntelligenceTheorem proving with abstraction
143 Citations1981David A. Plaisted
The use of abstractions permits the advantages of set-of-support strategies to be realized in arbitrary complete non set- of-support resolution strategies.
The Implications of Induction
118 Citations2019Larry Cohen
Machine LearningLearning Two-Tiered Descriptions of Flexible Concepts: The POSEIDON System
103 Citations1992Francesco Bergadano, Stan Matwin +2 more
The proposed method has been implemented in the POSEIDON system, and experimentally tested on two real-world problems: learning the concept of an acceptable union contract, and learning voting patterns of Republicans and Democrats in the U.S. Congress.
Journal of the Learning SciencesA Theory of Questions and Question Asking
103 Citations1991Ashwin Ram
A theory of questions and of question asking is developed, motivated both by cognitive and computational considerations, and it is discussed the theory in the context of story understanding.
Machine LearningRESEARCH IN MACHINE LEARNING
82 Citations1990Ryszard S. Michalski, Yves Kodratoff
This chapter gives a brief account of the recent progress and prospective research directions in the field, attempts to clarify some basic concepts, proposes a multicriteria classification of learning methods, and concludes with a brief description of each chapter.
U-M-I Out-of-Print Books on DemandElements of logic
60 Citations1932Charles S. Peirce, Charles Hartshorne +1 more
Elsevier eBooksAcquiring Effective Search Control Rules: Explanation-Based Learning in the PRODIGY System
57 Citations1987Steven Minton, Jaime Carbonell +3 more
Three mechanisms incorporated in the PRODIGY system for attacking the problem of effective control knowledge are described, including one that is selective about what it learns from a particular example, and one that empirically evaluates the utility of the rules it learns.
International Joint Conference on Artificial IntelligenceThe use of explanations for similarity-based learning
51 Citations1987Andrea Danyluk
An integrated learning approach that combines Explanation-Based and Similarity-Based Learning methods to make learning in an inherently complex domain feasible is proposed.
Applied IntelligenceThe use of explicit goals for knowledge to guide inference and learning
44 Citations1992Ashwin Ram, Lawrence Hunter
A theory of knowledge goals, or desires for knowledge, and their use in the processes of understanding and learning is presented, illustrated using a natural language understanding program that learns by reading novel or unusual newspaper stories, and a differential diagnosis program that improves its accuracy with experience.
A Strategy for Constructing New Predicates in First-Order Logic.
37 Citations1988Stephen Muggleton
An enhancement to CIGOL's learning strategy is described which strongly constrains the formation of new concepts and hypotheses and its improved performance is demonstrated by way of the automatic discovery of the concept of radiation.
Elsevier eBooksLIMITATIONS ON INDUCTIVE LEARNING
34 Citations1989Thomas G. Dietterich
This paper explores the proposition that inductive learning from examples is fundamentally limited to learning only a small fraction of the total space of possible hypotheses and proves an upper bound on the maximum number of concepts reliably learnable from m training examples.
Toward a Unified Theory of Learning: Multistrategy Task-adaptive Learning
32 Citations1993Ryszard S. Michalski
Integrating Explanation-Based and Empirical Learning Methods in OCCAM.
32 Citations1988Michael J. Pazzani
This paper focuses on OCCAM, a program that has the capability to acquire via empirical means the knowledge needed for analytical learning through the use of empirical techniques to acquire a domain theory for explanation based learning.
International Joint Conference on Artificial IntelligenceDISCIPLE-1: interactive apprentice system in weak theory fields
29 Citations1987Yves Kodratoff, Gheorghe Tecuci
A first prototypical version of DISCIPLE and its use to the design of technologies for the manufacturing of loudspeakers is presented and demonstrates not only that over-generalization is harmless but also useful and necessary, when interacting with a user.
Comparing Symbolic and Subsymbolic Learning: Three Studies
26 Citations1992Janusz Wnek, Ryszard S. Michalski
Kluwer international series in engineering and computer scienceOverview of the Odysseus Learning Apprentice
24 Citations1986David C. Wilkins, William J. Clancey +1 more
An apprentice learning system under development is described that allows an expert system to use some of these same methods that aid a expert system in transferring expertise to and from its knowledge base.
Machine LearningPlausible justification trees: A framework for deep and dynamic integration of learning strategies
23 Citations1993Gheorghe Tecuci
A concept-learning and theory-revision method that was developed in this framework that allows the system to learn from one or from several (positive and/or negative) examples, and to both generalize and specialize its knowledge base.
Elsevier eBooksA Method for Multistrategy Task-adaptive Learning Based on Plausible Justifications
20 Citations1991Gheorghe Tecuci, Ryszard S. Michalski
The paper presents a MTL method which is based on building a plausible justification that the learner's input is a consequence of its BK and integrates dynamically a whole range of learning strategies.
Lecture notes in computer scienceInput understanding as a basis for multistrategy task-adaptive learning
13 Citations1991Gheorghe Tecuci, Ryszard S. Michalski
The paper explores several general issues in developing a multistrategy task-adaptive learning (MTL) system, which aims at integrating a whole range of learning strategies, such as explanation-based learning, empirical generalization, abduction, constructive induction, learning by analogy and abstraction.
Elsevier eBooksAbstracting Concepts with Inverse Resolution
12 Citations1991Attilio Giordana, Lorenza Saitta +1 more
A restricted class of semantic abstraction (CP-abstraction) is characterized, having the property of preserving the concept instances with respect to a given model and fits into the paradigm of inverse resolution, which has been proposed as a framework for constructive learning.
Toward a Unified Theory of Learning: An Outline of Basic Ideas
9 Citations1991Ryszard S. Michalski
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.
Elsevier eBooksDECLARATIVE BIAS FOR STRUCTURAL DOMAINS
8 Citations1989Benjamin N. Grosof, Stuart Russell
The solution extends the declarative approach to bias of (Russell and Grosof, 1987) by formalizing the beliefs about relevancy in a more complex form that expresses the preservation of properties under mappings, using second-order logic to express the existence of isomorphisms.