A historical perspective on Machine Learning (on the occasion of - PowerPoint PPT Presentation

A historical perspective on Machine Learning (on the occasion of the 25th Benelearn) Luc De Raedt

A historical perspective on Machine Learning (on the occasion of the 25th Benelearn) Luc De Raedt

A historical perspective on Machine Learning (on the occasion of the 25th Benelearn) Luc De Raedt W ARNING ! Based on a true story of Machine Learning

A historical perspective on Machine Learning (on the occasion of the 25th Benelearn) Luc De Raedt S C I E N T I F I C W ARNING ! of Machine Learning ADVISOR Y Based on a true story P E R S O N A L P E R S P E C T I V E

Machine Learning: an AI approach

Machine Learning: an AI approach Ryszard Michalski …

Machine Learning: an AI approach Ryszard Michalski, Tom Mitchell, Jaime Carbonell

Machine Learning: an AI approach Ryszard Michalski, Tom Mitchell, Jaime Carbonell 1986 1990 1994 1983

1983 Preface

Before 1980 —Handbook of AI 1981 overview https://archive.org/details/handbookofartific02barr

Before 1980 —Handbook of AI 1981 overview https://archive.org/details/handbookofartific02barr

Menace (Michie 63)

O X O O X O X X X to move X Choose box corresponding to current state Execute move Choose pearl at random from box

Menace (Michie 1963) O X O Learns Tic-Tac-Toe Hardware: X 287 Boxes 
 (1 for each state) Pearls in 9 colors 
 (1 color per position) Play principle: Choose box corresponding to current state Choose pearl at random from box Play corresponding move Learning algorithm: Game lost -> retain all pearls used 
 ( negative reword - reinforcement) Game won -> for each select pearl, add a pearl of the same color to box ( positive reward - reinforcement )

BOXES (1968) https://www.youtube.com/watch?v=qF2fFMrNUCQ • basis of reinforcement learning

CONTENTS v Preface 1 PART ONE GENERAL ISSUES IN MACHINE LEARNING Chapter 1 An Overview of Machine Learning 3 Jaime G. Carbonell, Ryszard S. Michalski, and Tom M. Mitchell 1.1 Introduction 3 1.2 3 The Objectives of Machine Learning 1.3 7 A Taxonomy of Machine Learning Research 1.4 An Historical Sketch of Machine Learning 14 16 1.5 A Brief Reader's Guide Chapter 2 25 Why Should Machines Learn? Herbert A. Simon 2.1 25 Introduction 2.2 25 Human Learning and Machine Learning 2.3 28 What is Learning? 2.4 30 Some Learning Programs 2.5 32 Growth of Knowledge in Large Systems 2.6 34 A Role for Learning 2.7 35 Concluding Remarks PART TWO 39 LEARNING FROM EXAMPLES Chapter 3 A Comparative Review of Selected Methods 41 for Learning from Examples Thomas G. Dietterich and Ryszard S. Michalski 3.1 41 Introduction 3.2 49 Comparative Review of Selected Methods

Why should machines learn ? Herbert Simon (1916-2001) Turing Award 1975, Nobel prize Economics 1978

viii CONTENTS 3.3 Conclusion 75 ix CONTENTS Chapter 4 A Theory and Methodology of Inductive 83 Learning 217 7.5 Summary of Geometry Learning Ryszard S. Michalski 221 Chapter 8 Using Proofs and Refutations to Learn from 4.1 Introduction 83 Experience 4.2 Types of Inductive Learning .87 Frederick Hayes-Roth 4.3 Description Language 94 8.1 Introduction 221 4.4 Problem Background Knowledge 96 222 8.2 The Learning Cycle 4.5 Generalization Rules 103 225 8.3 Five Heuristics for Rectifying Refuted Theories 112 4.6 The Star Methodology 8.4 Computational Problems and Implementation 234 116 4.7 An Example Techniques 4.8 Conclusion 123 8.5 Conclusions 238 4.A Annotated Predicate Calculus (APC) 130 LEARNING FROM OBSERVATION AND 241 PART FOUR PART THREE LEARNING IN PROBLEM-SOLVING AND 135 DISCOVERY PLANNING 243 Chapter 9 The Role of Heuristics in Learning by Chapter 5 Learning by Analogy: Formulating and 137 Discovery: Three Case Studies Generalizing Plans from Past Experience Douglas B. Lenat Jaime G. Carbonell 243 9.1 Motivation 137 5.1 Introduction 9.2 Overview 245 139 5.2 Problem-Solving by Analogy 9.3 Case Study 1 : The AM Program; Heuristics 249 149 5.3 Evaluating the Analogical Reasoning Process Used to Develop New Knowledge 5.4 Learning Generalized Plans 151 9.4 A Theory of Heuristics 263 5.5 Concluding Remark 159 276 9.5 Case Study 2: The Eurisko Program; Heuristics Chapter 6 Learning by Experimentation: Acquiring and 163 Used to Develop New Heuristics Refining Problem-Solving Heuristics 282 9.6 Heuristics Used to Develop New Tom M. Mitchell, Paul E. Utgoff, and Ranan Banerji Representations 286 9.7 Case Study 3: Biological Evolution; Heuristics 6.1 Introduction 163 Used to Generate Plausible Mutations 6.2 The Problem 164 302 9.8 Conclusions 6.3 Design of LEX 167 6.4 New Directions: Adding Knowledge to Augment 180 Rediscovering Chemistry With the BACON 307 Chapter 10 Learning System 6.5 Summary 189 Pat Langley, Gary L. Bradshaw, and Herbert A. Simon Chapter 7 Acquisition of Proof Skills in Geometry 191 307 John R. Anderson 10.1 Introduction 10.2 An Overview of BACON.4 309 7.1 Introduction 191 312 10.3 The Discoveries of SACON.4 7.2 A Model of the Skill Underlying Proof Generation 193 319 10.4 Rediscovering Nineteenth Century Chemistry 7.3 Learning 201 10.5 Conclusions 326 7.4 Knowledge Compilation 202

CONTENTS xi x CONTENTS Chapter 14 The Instructible Production System: A 429 Chapter 11 Retrospective Analysis Learning From Observation: Conceptual 331 Clustering Michael D. Rychener Ryszard S. Michalski and Robert E. Stepp 430 14.1 The Instructible Production System Project 11.1 Introduction 332 14.2 Essential Functional Components of Instructible 436 11.2 Conceptual Cohesiveness 333 Systems 11.3 Terminology and Basic Operations of the 336 14.3 Survey of Approaches 443 Algorithm 14.4 Discussion 453 11.4 A Criterion of Clustering Quality 344 461 PART SIX APPLIED LEARNING SYSTEMS 11.5 Method and Implementation 345 11.6 An Example of a Practical Problem: Constructing 358 Chapter 15 Learning Efficient Classification Procedures 463 a Classification Hierarchy of Spanish Folk Songs and their Application to Chess End Games 11.7 Summary and Some Suggested Extensions of 360 J. Ross Quinlan the Method 15.1 Introduction 463 PART FIVE LEARNING FROM INSTRUCTION 365 465 15.2 The Inductive Inference Machinery 15.3 The Lost N-ply Experiments 470 Chapter 12 Machine Transformation of Advice into a 367 15.4 Approximate Classification Rules 474 Heuristic Search Procedure 15.5 Some Thoughts on Discovering Attributes 477 David Jack Mostow 15.6 Conclusion 481 12.1 Introduction 367 Chapter 16 Inferring Student Models for Intelligent 483 12.2 Kinds of Knowledge Used 370 Computer-Aided Instruction 12.3 A Slightly Non-Standard Definition of Heuristic 374 Derek H. Sleeman Search 12.4 Instantiating the HSM Schema for a Given 378 483 16.1 Introduction Problem 16.2 Generating a Complete and Non-redundant Set 488 12.5 Refining HSM by Moving Constraints Between 384 of Models Control Components 16.3 Processing Domain Knowledge 503 12.6 Evaluation of Generality 398 16.4 Summary 507 12.7 Conclusion 399 16.A An Example of the SELECTIVE Algorithm: 510 12.A Index of Rules 403 LMS-I's Model Generation Algorithm 405 Chapter 13 Learning by Being Told: Acquiring Comprehensive Bibliography of Machine Learning 511 Knowledge for Information Management Paul E. Utgoff and Bernard Nudel Norm Haas and Gary G. Hendrix Glossary of Selected Terms In Machine Learning 551 13.1 Overview 405 13.2 Technical Approach: Experiments with the 408 About the Authors 557 KLAUS Concept 13.3 More Technical Details 413 Author Index 563 13.4 Conclusions and Directions for Future Work 418 13.A Training NANOKLAUS About Aircraft Carriers 422 Subject Index 567

1980 … 1986 • First workshops on Machine Learning (first conference in 1993) • Focus on AI and Cognitive Science paradigm • Focus on SYMBOLIC Methods, on HUMAN like learning, on AUTOMATED DISCOVERY • IJCAI 85 in LA had 3000 academic participants (10 000 with industry included?) These were the days of expert systems • No role for SUBSYMBOLIC methods / NEURAL NETS • NIPS would start in 1986, with the revival of Neural Networks (Parallel Distributed Processing / Connectionism — Rumelhart and McClelland) • https://www.youtube.com/watch?v=ilP4aPDTBPE (1989)