Systems that learn : an introduction to learning theory / Sanjay Jain ... [and others].

Format:

Book

Author/Creator:

Jain, Sanjay, 1965 February 22- author.

Series:

Learning, development, and conceptual change

Language:

English

Subjects (All):

Human information processing--Mathematical models.

Learning--Mathematical models.

Learning, Psychology of.

Human information processing.

Physical Description:

1 online resource (xii, 317 pages) : illustrations.

Edition:

Second edition.

Other Title:

MIT Press CogNet.

Place of Publication:

Cambridge, Massachusetts : The MIT Press, [1999]

System Details:

text file

Summary:

Formal learning theory is one of several mathematical approaches to the study of intelligent adaptation to the environment. The analysis developed in this book is based on a number theoretical approach to learning and uses the tools of recursive-function theory to understand how learners come to an accurate view of reality. This revised and expanded edition of a successful text provides a comprehensive, self-contained introduction to the concepts and techniques of the theory. Exercises throughout the text provide experience in the use of computational arguments to prove facts about learning.

Contents:

1.1 Empirical inquiry 3

1.2 Paradigms 5

1.3 Some simple paradigms 5

2 Formalities 15

3 Identification 27

3.1 Languages as theoretically possible realities 27

3.2 Language identification: Hypotheses, data 31

3.3 Language identification: Scientists 32

3.4 Language identification: Scientific success 35

3.5 Identification as a limiting process 38

3.6 Characterization of identifiable 40

3.7 Some alternative paradigms 44

3.8 Memory-limited scientists 45

3.9 Second paradigm: Identification of functions 48

3.10 Characterization of identifiable 51

3.12 Exercises 55

4 Identification by Computable Scientists 61

4.2 Language identification by computable scientist 63

4.3 Function identification by computable scientist 69

4.4 Parameterized scientists 75

4.5 Exact identification 81

II Fundamental Paradigms Generalized 89

5 Strategies for Learning 91

5.1 Strategies for language identification: Introduction 91

5.2 Constraints on potential conjectures 92

5.3 Constraints on the use of information 99

5.4 Constraint on convergence 100

5.5 Constraints on the relation between conjectures 107

5.6 Strategies for function identification 115

6 Criteria of Learning 127

6.1 Criteria for function identification 127

6.2 Criteria of language identification 138

7 Inference of Approximations 151

7.1 Approximations 151

7.3 Approximate explanatory identification 154

7.4 Uniform approximate explanatory identification 158

8 Environments 167

8.1 Inaccurate data 168

8.2 Texts with additional structure 180

8.3 Multiple texts 184

III Part III: Additional Topics 195

9 Team and Probabilistic Learning 197

9.2 Motivation for identification by teams 197

9.3 Team identification of functions 199

9.4 Identification by probabilistic scientists 201

9.5 Team Ex-identification 212

9.6 Team and probabilistic identification of languages 213

10 Learning with Additional Information 221

10.2 Upper bound on the size of hypothesis 222

10.3 Approximate hypotheses as additional information 235

11 Learning with Oracles 251

11.2 Oracle scientists 251

11.3 Function identification by oracle scientists 252

11.4 Language identification by oracle scientists 257

12 Complexity Issues in Identification 261

12.2 Mind change complexity 262

12.3 Number of examples required 264

12.4 An axiomatic approach to complexity of convergence 267

12.5 Strictly minimal identification of languages 268

12.6 Nearly minimal identification 273

13 Beyond Identification by Enumeration 281

13.1 Gold's and Barzdins conjectures 281

13.2 Fulk's refutation of Barzdins conjecture 282.

Notes:

"A Bradford book."

Includes bibliographical references (pages 289-302) and indexes.

Description based on print version record.

Other Format:

Print version: Systems that learn.

ISBN:

9780262276252

0262276259

Access Restriction:

Restricted for use by site license.