Skip to main content
Fachgebiete
Automobil + Motoren Bauwesen + Immobilien Business IT + Informatik Elektrotechnik + Elektronik Energie + Nachhaltigkeit Finance + Banking Management + Führung Marketing + Vertrieb Maschinenbau + Werkstoffe Versicherung + Risiko
DE
EN
Bücher
Zeitschriften
Themenseiten
Organisationspsychologie Projektmanagement Marketing Smart Manufacturing
Weitere Formate
Podcasts Webinare Technik Webinare Wirtschaft Kongresse Veranstaltungskalender Awards
Jetzt Einzelzugang starten
Zugang für Unternehmen
Referenzkunden
SLX-Digitalkonferenz: Zukunftswerkstatt 2024

Mehr Umsatz mit Top-Kontakten

Am 7. Mai erfahren Sie, wie Vertriebsteams Leadgenerierung, Leadnurturing und Leadstrategien effizient steuern können.
Erweiterte Suche
Anmelden
nach oben

1983 | Buch

An Overview of Machine Learning Kapitel lesen Erstes Kapitel lesen

Machine Learning

An Artificial Intelligence Approach

herausgegeben von: Ryszard S. Michalski, Jaime G. Carbonell, Tom M. Mitchell

Verlag: Springer Berlin Heidelberg

Buchreihe : Symbolic Computation

Enthalten in: Professional Book Archive

Einloggen, um Zugang zu erhalten
insite
SUCHEN

Über dieses Buch

The ability to learn is one of the most fundamental attributes of intelligent behavior. Consequently, progress in the theory and computer modeling of learn­ ing processes is of great significance to fields concerned with understanding in­ telligence. Such fields include cognitive science, artificial intelligence, infor­ mation science, pattern recognition, psychology, education, epistemology, philosophy, and related disciplines. The recent observance of the silver anniversary of artificial intelligence has been heralded by a surge of interest in machine learning-both in building models of human learning and in understanding how machines might be endowed with the ability to learn. This renewed interest has spawned many new research projects and resulted in an increase in related scientific activities. In the summer of 1980, the First Machine Learning Workshop was held at Carnegie-Mellon University in Pittsburgh. In the same year, three consecutive issues of the Inter­ national Journal of Policy Analysis and Information Systems were specially devoted to machine learning (No. 2, 3 and 4, 1980). In the spring of 1981, a special issue of the SIGART Newsletter No. 76 reviewed current research projects in the field. . This book contains tutorial overviews and research papers representative of contemporary trends in the area of machine learning as viewed from an artificial intelligence perspective. As the first available text on this subject, it is intended to fulfill several needs.

Inhaltsverzeichnis

Frontmatter

General Issues in Machine Learning

Frontmatter
Chapter 1. An Overview of Machine Learning
Abstract
Learning is a many-faceted phenomenon. Learning processes include the acquisition of new declarative knowledge, the development of motor and cognitive skills through instruction or practice, the organization of new knowledge into general, effective representations, and the discovery of new facts and theories through observation and experimentation. Since the inception of the computer era, researchers have been striving to implant such capabilities in computers. Solving this problem has been, and remains, a most challenging and fascinating long-range goal in artificial intelligence (AI). The study and computer modeling of learning processes in their multiple manifestations constitutes the subject matter of machine learning.
Jaime G. Carbonell, Ryszard S. Michalski, Tom M. Mitchell
Chapter 2. Why Should Machines Learn?
Abstract
When I agreed to write this chapter, I thought I could simply expand a paper that I wrote for the Carnegie Symposium on Cognition, since the topic of that symposium was also learning. The difficulty with plagiarizing that paper is that it was really about psychology, whereas this book is concerned with machine learning. Now although we all believe machines can simulate human thought—unless we’re vitalists, and there aren’t any of those around any more—still, I didn’t think that was what was intended by the title of the book. I didn’t think it was appropriate to write about psychology.
Herbert A. Simon

Learning from Examples

Frontmatter
Chapter 3. A Comparative Review of Selected Methods for Learning from Examples
Abstract
Research in the area of learning structural descriptions from examples is reviewed, giving primary attention to methods of learning characteristic descriptions of single concepts. In particular, we examine methods for finding the maximally-specific conjunctive generalizations (MSC-generalizations) that cover all of the training examples of a given concept. Various important aspects of structural learning in general are examined, and several criteria for evaluating structural learning methods are presented. Briefly, these criteria include (i) adequacy of the representation language, (ii) generalization rules employed, (iii) computational efficiency, and (iv) flexibility and extensibility. Selected learning methods developed by Buchanan, et al., Hayes-Roth, Vere, Winston, and the authors are analyzed according to these criteria. Finally, some goals are suggested for future research.
Thomas G. Dietterich, Ryszard S. Michalski
Chapter 4. A Theory and Methodology of Inductive Learning
Abstract
The presented theory views inductive learning as a heuristic search through a space of symbolic descriptions, generated by an application of various inference rules to the initial observational statements. The inference rules include generalization rules, which perform generalizing transformations on descriptions, and conventional truth-preserving deductive rules. The application of the inference rules to descriptions is constrained by problem background knowledge, and guided by criteria evaluating the “quality” of generated inductive assertions.
Based on this theory, a general methodology for learning structural descriptions from examples, called Star, is described and illustrated by a problem from the area of conceptual data analysis.
Ryszard S. Michalski

Learning in Problem-Solving and Planning

Frontmatter
Chapter 5. Learning by Analogy: Formulating and Generalizing Plans from Past Experience
Abstract
Analogical reasoning is a powerful mechanism for exploiting past experience in planning and problem solving. This chapter outlines a theory of analogical problem solving based on an extension to means-ends analysis. An analogical transformation process is developed to extract knowledge from past successful problem-solving situations that bear a strong similarity to the current problem. Then, the investigation focuses on exploiting and extending the analogical reasoning model to generate useful exemplary solutions to related problems from which more general plans can be induced and refined. Starting with a general analogical inference engine, problem-solving experience is, in essence, compiled incrementally into effective procedures that solve various classes of problems in an increasingly reliable and direct manner.
Jaime G. Carbonell
Chapter 6. Learning by Experimentation: Acquiring and Refining Problem-Solving Heuristics
Abstract
This chapter concerns learning heuristic problem-solving strategies through experience. In particular, we focus on the issue of learning heuristics to guide a forward-search problem solver, and describe a computer program called lex, which acquires problem-solving heuristics in the domain of symbolic integration. lex acquires and modifies heuristics by iteratively applying the following process: (i) generate a practice problem; (ii) use available heuristics to solve this problem; (iii) analyze the search steps performed in obtaining the solution; and (iv) propose and refine new domain-specific heuristics to improve performance on subsequent problems. We describe the methods currently used by lex, analyze strengths and weaknesses of these methods, and discuss our current research toward more powerful approaches to learning heuristics.
Tom M. Mitchell, Paul E. Utgoff, Ranan Banerji
Chapter 7. Acquisition of Proof Skills in Geometry
Abstract
The ACT theory of learning is applied to the domain of high school geometry. The concern is with how students become skilled at planning a proof of a geometry problem. A general control structure is proposed for integrating backward and forward search in proof planning. This is embodied in a production system framework. Two types of learning are described. Knowledge compilation is concerned. with how students transit from a declarative characterization of the domain to a set of operators for performing a specific task in that domain. Tuning is concerned with how students learn which problem features are predictive of the success of which operators.
John R. Anderson
Chapter 8. Using Proofs and Refutations to Learn from Experience
Abstract
To learn, a learner needs to formulate plans, monitor the plan execution to detect violated expectations, and then diagnose and rectify errors which the dis-confirming data reveal. In this paper, five heuristic methods are presented for repairing flawed beliefs. These beliefs are considered as theories that predict effects of actions. Theories presuppose particular structural characteristics. When data disconfirm a theory, the heuristics proposed suggest specific ways to remedy the theory, including restricting the conditions for invoking the theory and weakening the theory’s predictions. The five methods accomplish retraction, exclusion, avoidance, assurance and inclusion of outcomes that disconfirm a theory’s predictions. Each proposed theory fix produces as a by-product new domain concepts that capture environmental characteristics of instrumental value to the learner. The techniques proposed here provide the first analytical methods for constructing new knowledge. They extend and make practical the ideas of proofs and refutations originally introduced by Lakatos.
Frederick Hayes-Roth

Learning from Observation and Discovery

Frontmatter
Chapter 9. The Role of Heuristics in Learning by Discovery: Three Case Studies
Abstract
As artificial intelligence (AI) programs are called upon to exhibit increasingly complex behaviors, their builders are faced with the growing task of inserting more and more knowledge into the machine. One long-range solution is for the program, by itself, to learn via discovery. The first case study presented, am, demonstrates that new domains of knowledge can be developed mechanically by using heuristics. Yet as new domain concepts, facts, and conjectures emerge, specific new heuristics, or informal judgmental rules, are needed. They in turn can be discovered by using a body of heuristics for guidance. The second case study, eurisko, has already achieved some promising results in this endeavor. If this process—using heuristics to guide “learning by discovery”—is so powerful and simple, one wonders why, for instance, nature has not adopted an analogous mechanism to guide evolution. Indeed, the final part of the article is a speculation that evolution does function in that manner. In place of the conventional Darwinian process of random mutation, we hypothesize a more powerful plausible generation scheme.
Douglas B. Lenat
Chapter 10. Rediscovering Chemistry with the Bacon System
Abstract
BACON.4 is a production system that discovers empirical laws. The program represents information at varying levels of description, with higher levels summarizing the levels below them. BACON.4 employs a small set of data-driven heuristics to detect regularities in numeric and nominal data. These heuristics note constancies and trends, causing BACON.4 to formulate hypotheses, to define theoretical terms, and to postulate intrinsic properties. The introduction of intrinsic properties plays an important role in BACON.4’s rediscovery of Ohm’s law for electric circuits and Archimedes’ law of displacement. When augmented with a heuristic for noting common divisors, the system is able to replicate a number of early chemical discoveries, arriving at Proust’s law of definite proportions, Gay-Lussac’s law of combining volumes, Cannizzaro’s determination of the relative atomic weights, and Prout’s hypothesis. The BACON.4 heuristics, including the new technique for finding common divisors, appear to be general mechanisms applicable to discovery in diverse domains.
Pat Langley, Gary L. Bradshaw, Herbert A. Simon
Chapter 11. Learning from Observation: Conceptual Clustering
Abstract
An important form of learning from observation is constructing a classification of given objects or situations. Traditional techniques for this purpose, developed in cluster analysis and numerical taxonomy, are often inadequate because they arrange objects into classes solely on the basis of a numerical measure of object similarity. Such a measure is a function only of compared objects and does not take into consideration any global properties or concepts characterizing object classes. Consequently, the obtained classes may have no simple conceptual description and may be difficult to interpret.
The above limitation is overcome by an approach called conceptual clustering, in which a configuration of objects forms a class only if it is describable by a concept from a predefined concept class. This chapter gives a tutorial overview of conjunctive conceptual clustering, in which the predefined concept class consists of conjunctive statements involving relations on selected object attributes. The presented method arranges objects into a hierarchy of classes closely circumscribed by such conjunctive descriptions. Descriptions stemming from each node are logically disjoint, satisfy given background knowledge, and optimize a certain global criterion.
The method is illustrated by an example in which the conjunctive conceptual clustering program CLUSTER/2 constructed a classification hierarchy of a large collection of Spanish folk songs. The conclusion suggests some extensions of the method and topics for further research.
Ryszard S. Michalski, Robert E. Stepp

Learning from Instruction

Frontmatter
Chapter 12. Machine Transformation of Advice Into a Heuristic Search Procedure
Abstract
A key problem in learning by being told is operationalization: the development of procedures to implement advice that is not directly executable by the learner, such as the advice “avoid taking points” in the card game hearts. One way to operationalize such advice is to reformulate it in terms of a general “weak method”, such as heuristic search. This chapter is a case study in the mechanical mapping of domain-specific problems onto general methods, using as a detailed example the derivation of a heuristic search procedure for the advice “avoid taking points.” The derivation consists of a series of problem transformations leading from the advice statement to an executable procedure. The operators used to perform these transformations are implemented in a program called FOO as domain-independent transformation rules that access a knowledge base of task domain concepts. Some of the rules construct a crude generate-andtest procedure; others improve it by deriving new heuristics based on domain knowledge and problem analysis. To test its generality, FOO was also used to operationalize a music composition task; many of the same rules proved applicable.
David Jack Mostow
Chapter 13. Learning By Being Told: Acquiring Knowledge for Information Management
Abstract
This chapter discusses machine-learning aspects of a project whose broad goal is to create computer systems that can aid users in managing information. The specific learning problem discussed is how to enable computer systems to acquire information about domains with which they are unfamiliar from people who are experts in those domains, but have little or no training in computer science. The information to be acquired is that needed to support question-answering or fact-retrieval tasks, and the type of learning to be employed is learning by being told.
Norman Haas, Gary G. Hendrix
Chapter 14. The Instructible Production System: A Retrospective Analysis
Abstract
In building systems that acquire knowledge from tutorial instruction, progress depends on determining certain functional requirements and ways for them to be met. The Instructible Production System (IPS) project has explored learning by building a series of experimental systems. These systems can be viewed as being designed to explore the satisfaction of some of the requirements, both by basic production system mechanisms and by features explicitly programmed as rules. The explorations have brought out the importance of considering in advance (as part of the kernel design) certain functional components rather than having them be filled in by instruction. The need for the following functional components has been recognized:
  • interaction language
  • organization of procedural elements
  • explanation of system behavior
  • accommodation to new knowledge
  • connection of goals with system capabilities
  • reformulation (mapping) of knowledge
  • evaluation of behavior
  • compilation to achieve efficiency and automaticity
Since the experimental systems have varied in their effectiveness, some general conclusions can be drawn about the relative merits of various approaches. Seven such approaches are discussed here, with particular attention to the three whose behavior can be most effectively compared, and which reflect the temporal development of the project.
Michael D. Rychener

Applied Learning Systems

Frontmatter
Chapter 15. Learning Efficient Classification Procedures and Their Application to Chess End Games
Abstract
A series of experiments dealing with the discovery of efficient classification procedures from large numbers of examples is described, with a case study from the chess end game king-rook versus king-knight. After an outline of the inductive inference machinery used, the paper reports on trials leading to correct and very fast attribute-based rules for the relations lost 2-ply and lost 3-ply. On another tack, a model of the performance of an idealized induction system is developed and its somewhat surprising predictions compared with observed results. The paper ends with a description of preliminary work on the automatic specification of relevant attributes.
J. Ross Quinlan
Chapter 16. Inferring Student Models for Intelligent Computer-Aided Instruction
Abstract
The ability to formulate students’ models automatically is a critical component of intelligent teaching systems. This chapter reports current progress on inducing models of simple algebraic skills from observed student performance in the context of the Leeds Modeling System (LMS). A model consists of an ordered production system with potentially incorrect variants of some rules (called mal-rules). Constraining the number of plausible models that account for the observed student problem-solving behavior has proved to be a major undertaking. First, the current rule-based formulation is presented. Its shortcomings are indicated and a revised analysis is given which is demonstrated to create a complete and non-redundant set of models. Second, the related issue of generating a minimal problem template is discussed. Such a template represents the simplest type of problem that insures all rules in a model are exercised. Finally, the significance of this type of analysis for other areas of AI is indicated.
Derek H. Sleeman
Backmatter
Metadaten
Titel
Machine Learning
herausgegeben von
Ryszard S. Michalski
Jaime G. Carbonell
Tom M. Mitchell
Copyright-Jahr
1983
Verlag
Springer Berlin Heidelberg
Electronic ISBN
978-3-662-12405-5
Print ISBN
978-3-662-12407-9
DOI
https://doi.org/10.1007/978-3-662-12405-5