Artificial Intelligence Principles and Practice (George F. Luger) (z-library.sk, 1lib.sk, z-lib.sk)

Artificial Intelligence: Principles and Practice George F. Luger

Artificial Intelligence: Principles and Practice

George F. Luger Artificial Intelligence: Principles and Practice

George F. Luger University of New Mexico Albuquerque, NM, USA ISBN 978-3-031-57436-8 ISBN 978-3-031- 57437-5 (eBook) https://doi.org/10.1007/978-3-031-57437-5 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2025 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, spe- cifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, com- puter software, or by similar or dissimilar methodology now known or hereaf- ter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a war- ranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neu- tral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzer- land If disposing of this product, please recycle the paper.

For Kate V

Preface This book offers a comprehensive introduction to the exciting and too often mysterious discipline of Artificial Intelligence. As an introduction to AI, it covers the foundational computa- tional technologies that have supported work in AI since its inception over 70 years ago. In Part I, it covers the mathemati- cal, philosophical, and engineering disciplines that make artifi- cial intelligence possible. AI is not a standalone discipline but rather integrates and expands the many insights of the psychol- ogists, mathematicians, neuroscientists, and others that are part of its current research and practice. The final chapters, Part VIII, focus on the limitations, ethical issues, and future promise of AI. Here we present many of the social issues that must be considered when integrating AI technology into our modern world. There are three major design paradigms that have emerged in AI research and engineering over the past 70 years. The first paradigm, begun in the 1950s, is the symbol-based approach to problem-solving. Symbol-based AI is founded on graph the- ory, search algorithms, and the use of representational tech- niques including semantic networks and the propositional and predicate calculi. These technologies are covered in Parts II through V. The second paradigm of AI practice is neural or connection- ist network problem-solving. Although this tradition had its roots in the 1940s and was mentioned in the AI manifesto made at the Dartmouth summer workshop in 1956, it did not reach its full potential until the creation of the Boltzmann machine and the backpropagation algorithms of the 1970s and 1980s (Werbos 1994; Rumelhart and McClelland 1986). Since then, neural networks and deep learning have produced many quality results in image processing, human language analysis, and other areas of complex information processing. Neural net- work problem-solving, with an introduction to deep learning, is covered in the six chapters of Part VI. Probabilistic or stochastic approaches to problem-solving make up the third major paradigm of AI. This approach uses Bayesian-based probability measures to address the uncertain- ties and impreciseness of world situations. Probabilistic AI began in the digit and image recognition studies of the 1950s, but only reached large-scale acceptance with successes in human language processing of the 1990s. Pearl’s (1988, 2000) insights, seen in the creation of Bayesian belief networks and dynamic Bayesian networks, allowed the many forms of Bayes- ian reasoning to be computationally tractable, more transpar- ent, and accountable to the user. Probabilistic models are presented in Part VII.

There are other minor paradigms, with much more limited use in current AI practice. These include genetic algorithms, artificial life, fuzzy and other logic-based forms for uncertain reasoning. Genetic algorithms, suggested by Holland in the 1970s, and emergent computation, first proposed by von Neu- mann in the 1940s, reflect Darwin’s insights on population growth and survival of the fittest members of a society. These techniques are presented in subsections of Part II. Fuzzy and other logic-based schemes for reasoning in uncertain situations are presented in 7 Chap. 12. The principles and practice of these paradigms make up the foundation and successes of work in current artificial intelli- gence. They also reflect the structure and the recommended uses of our course materials. 1.1 Using This Book This book is intended to be a general introduction to the tools and techniques of modern artificial intelligence. For an AI practitioner to have a full vision on the powers and promise of AI, they need a perspective that includes the intellectual roots, the evolving tool building, and an historical appreciation of our discipline. Further, as AI is intended to become an integral component of modern society, its developers and users must have an ethical stance and appreciation of its powers and limi- tations. Our book is designed to address these needs. This book has a modular design so that an instructor, or students exploring on their own, can get a focused view of the different components of modern AI.  The Table of Contents reflects the eight modules that make up the introduction to AI, and each module introduces that topic from its foundations. For example, Part II, on symbol-based AI, introduces graph theory with basic search algorithms. Similarly, Part VII pres- ents probabilistic AI, beginning with its mathematical justifica- tions. Using this modular perspective, we trust instructors and students will acquire only those components of the book that reflect the topics they wish to cover. Finally, Part I reviews the mathematical, philosophical, and engineering technology that makes AI possible, and Part VIII addresses topics supporting the ethical use of artificial intelli- gence in practice. The topics covered in these opening and clos- ing components of this book are critical considerations for the mature student, for the AI community itself, as well as for soci- ety at large. VII Preface

1.2 Programming » What we must learn to do we learn by doing… ARISTOTLE, Ethics AI practitioners come to understand their world, and in that process to appreciate intelligent problem solutions, through writing programs that solve important problems in that world. As we state many times throughout our book, a critical feature of artificial intelligence is the exploratory programming meth- odology. We come to understand and intelligently interact with our world by writing programs that function within it. With exploratory design we grow these programs by systematic approximation: as we see our programs failing to meet our requirements, we analyze them carefully and rebuild them, intending a better result. Thus, programming becomes a critical component of the AI toolkit. There are many requirements for a suitable AI pro- gramming language, including transparency, expressiveness, and efficiency, properties we will discuss throughout this book. Although there are multiple languages that meet these require- ments, we have built many of the algorithms of this book in three: Prolog, Lisp, and Java. We also reference many AI appli- cations represented in Python. How these languages and our example code can be paired with the contents of this book will be described in the introductions to each of the major sections of the book. 1.3 About the Author I have taught, done research, and consulted with government and industry for more than five decades. Through these experi- ences, I have come to understand the important supporting principles and design practices of modern artificial intelligence. My undergraduate and graduate education at Gonzaga Uni- versity included mathematics, philosophy, and classical lan- guages. I began a graduate program in mathematics at the University of Notre Dame, supported by an NSF Sputnik scholarship. I received an MS in mathematics from ND in 1969 and began an interdisciplinary PhD program at the University of Pennsylvania that same year. My first AI class at Penn, using Nils Nilsson’s 1971 AI text, Problem-Solving Methods in Artificial Intelligence, was taught by Prof John W. Carr III in I971. I completed my PhD in 1973 writing a dissertation that modeled human puzzle solving strat- egies with the data structures and search techniques afforded by computing. My first AI research presentation was in the graduate student research presentations component of the International Joint Conference on Artificial Intelligence at VIII Preface

Stanford University in 1973. During my time at Penn, I was able to visit the research labs of Marvin Minsky, Seymour Pap- ert, and Terry Winograd at MIT. I also visited Carnegie Mellon University and attended several of Herbert Simon’s seminars and lectures. In 1974, I began a 4½-year postdoctoral research fellowship at the Department of Artificial Intelligence of the University of Edinburgh. I started my postdoc under Professor Bernard Meltzer and at the next funding cycle began work with Profes- sor Alan Bundy. Our research group, called MECHO (MECHanics Oracle) was a very early user of the Prolog pro- gramming language. We built a program that solved applied mathematics problems stated in English (Bundy et  al. 1979; Luger 1981). My colleagues at the University of Edinburgh included Bob Kowalski, David Warren, Fernando Pereira, Richard Young, David MacQueen, Martha Stone Palmer, T.G.R. Bower, Jen- nifer Wishart, Brendan McGonigle, and Tim O’Shea. Rod Burstall and Donald Michie were also on the faculty at that time. Visiting collaborators included Michael Arbib, George Lakoff, Alan Robinson, and Yorick Wilks. It was truly an excit- ing time, at the beginnings of the AI promise, to work at the University of Edinburgh. In 1979, I became an Associate Professor of Computer Sci- ence at the University of New Mexico. Peder Johnson, a Pro- fessor of Psychology, and I founded the Cognitive Science program at UNM in the mid-1980s. Professor of Linguistics Caroline Smith and I created the UNM program in Computa- tional Linguistics in the early 2000s. My PhD students include Chayan Chakrabarti, Michael Darling, Paul DePalma, Sunny Fugate, Ben Gordon, Kshanti Greene, Bill Klein, Joseph Lewis, Linda Means, Dan Pless, Roshan Rommohan, Nakita Sakanenko, Jim Skinner, Carl Stern, and Bill Stubblefield. Many of their research contributions can be found in this pub- lication. NATO, the British Royal Society, NASA, the Department of Defense, the Departments of Energy and Transportation, the NIH, and other agencies including the Smithsonian Institu- tion have supported my research. I have worked with the Los Alamos and Sandia National Laboratories and have received research funding through, and consulted for, numerous private companies. My most recent National Science Foundation- supported research was in building algorithms for probabilistic diagnostic reasoning. My graduate students and I have devel- oped stochastic models, mostly as an extended form of dynamic Bayesian networks, that modeled complex environments such as the production of electric power using sodium-cooled nuclear reactors, as presented in 7 Chap. 24. For further detail on my research and publications, see 7 https://www. cs. unm. edu/~luger/ or Google “Luger AI”. IX Preface

Acknowledgments The cover art is taken from an important recent paper that combines the symbol-based approach to prob- lem-solving with deep learning. The AlphaGeometry program takes complex geometry problems stated in English, translates them into a set of premises and a theorem to be proven, and then finds a proof. The top figure is the diagram of the original prob- lem, and the lower figure represents the generalized result. The text states the problem, the premises and goal statements, and the proof. Full details may be found in Trinh et al. (2024). The figures and algorithms in this book come from many sources. I began teaching the AI class at the University of New Mexico in the early 1980s. Many components of Parts II through IV came from these class presentations. I was a consul- tant and instructor for a professional education group, Learn- ing Tree International, beginning in 1983. I designed their artificial intelligence and neural network courses and created teaching materials for professional education through the 1990s. Many figures and examples in this book were first used in these teaching environments. In 1989, the first edition of my AI text, Artificial Intelli- gence: Structures and Strategies for Complex Problem Solving, with Dr. William A. Stubblefield as co-author, was published by Benjamin Cummings. The sixth edition was published by Pearson Education in 2009. With the help of four UNM col- leagues, my book, Cognitive Science: The Science of Intelligent Systems, was published by Academic Press in 1994. Many of the figures and examples in these books were created with the help of my graduate students at UNM, with the financial sup- port of the University of New Mexico, and by use of my own research funding. Part I, 7 Chaps. 1 and 2 introduces historical background for current AI, and Part VIII, addressing ethically related AI issues, have long been an important focus for me. As the bard has said, “if you don’t know where you are going, you will end up somewhere else.” We must understand the roots of the arti- ficial intelligence phenomena within the Western traditions of science, engineering, and philosophy. If we don’t know our own roots, it will be difficult to support continued growth, avoid problems, and most importantly, acknowledge the ethical responsibilities related to the development and use of AI tech- nology. Parts II through V, the symbol-based approach to artificial intelligence, reflect the earliest component of my own work in AI.  As just noted, the figures and examples come from my teaching at UNM and for Learning Tree International. Many of the figures and examples of Part VI, on connec- tionist AI, come from my book Cognitive Science. Carl Stern is co-author of 7 Chaps. 15, 16, 19, and 20; Chayan Chakrabarti is the co-author of 7 Chaps. 17 and 18. In Part VII, on proba- bilistic models and algorithms, my graduate students, espe- X Preface

cially Dan Pless, Roshan Rammohen, Chayan Chakrabarti, Nikita Sakhanenko, and Michael Darling are responsible for many of the figures and examples. I thank my Springer Nature editor, Paul Drougas, for his continued support and encouragement during the creation of this manuscript. Paul also supported my recent Springer publi- cation Knowing our World: An Artificial Intelligence Perspective (Luger 2021). Springer’s Jacob Shmulewitz was very helpful in preparing the figures for this book. My wife, Kate Luger, and friend, Dr. Zack Bruce, have made this presentation more coherent with their careful proofreading. I am grateful for the research and publication assistance over the years of my many AI colleagues, students, and friends. My life has been enriched through their continuing friendships. George F. Luger Albuquerque, NM, USA 1 December 2024 XI Preface

Contents I Introducing Artificial Intelligence 1 The Pre-History of Artificial Intelligence ............................................................................................... 3 1.1 Mary Shelley, Frankenstein, and Prometheus ................................................................................................... 4 1.2 The Age of Rationalism ............................................................................................................................................. 6 1.3 The Empiricist Tradition ............................................................................................................................................ 8 1.4 Immanuel Kant: Bridging the Rationalist/Empiricist Viewpoints ............................................................... 9 1.5 The Reverend Thomas Bayes ................................................................................................................................... 11 1.6 The Mathematical Foundations for Artificial Intelligence ............................................................................. 11 1.7 American Pragmatism ............................................................................................................................................... 16 1.8 The Turing Test: Can a Machine Be “Intelligent”? ............................................................................................. 17 1.9 The 1956 Dartmouth Summer Workshop ........................................................................................................... 20 1.10 Summary ........................................................................................................................................................................ 24 1.11 Exercises ......................................................................................................................................................................... 24 2 Computing, Representations, and Definitions of Artificial Intelligence .......................... 27 2.1 Artificial Intelligence: Attempting a Definition ................................................................................................. 28 2.2 Computer-Based Representations of the World ............................................................................................... 30 2.3 The General Themes of Current AI Practice ........................................................................................................ 40 2.4 Summary of Part I and an Introduction to Part II ............................................................................................. 48 2.5 Exercises ......................................................................................................................................................................... 49 II Symbol-Based AI: Foundations 3 The State Space, Finite State Machines, and Artificial Life........................................................ 53 3.1 Graph Theory: The Origins of the State Space Representation ................................................................... 54 3.2 The State Space Representation ............................................................................................................................ 58 3.3 The Finite State Machine .......................................................................................................................................... 64 3.4 Artificial Life: The Emergence of Complexity ..................................................................................................... 67 3.5 Summary ........................................................................................................................................................................ 71 3.6 Exercises ......................................................................................................................................................................... 72 4 Searching the State Space............................................................................................................................... 73 4.1 Data-Driven and Goal-Driven Search ................................................................................................................... 74 4.2 Implementing Graph Search with the Backtrack Algorithm ........................................................................ 77 4.3 Breadth-First and Depth-First Search ................................................................................................................... 80 4.4 Extending Search Strategies to and/or Graphs ................................................................................................. 88 4.5 Summary ........................................................................................................................................................................ 94 4.6 Exercises ......................................................................................................................................................................... 95 5 Heuristic Search ..................................................................................................................................................... 97 5.1 An Introduction to Heuristics .................................................................................................................................. 98 5.2 Hill-Climbing ................................................................................................................................................................. 101 5.3 Best-First Search .......................................................................................................................................................... 104 XIII

5.4 Genetic Algorithms: Evaluating Multiple “Best” States .................................................................................. 113 5.5 In Summary ................................................................................................................................................................... 121 5.6 Exercises ......................................................................................................................................................................... 122 6 Heuristics: 2-Person Games and Theoretical Constraints .......................................................... 125 6.1 The Minimax Procedure on Exhaustively Searchable Graphs ...................................................................... 126 6.2 Using Minimax to a Fixed Ply Depth ..................................................................................................................... 128 6.3 The Alpha-Beta Procedure ....................................................................................................................................... 133 6.4 Multi-Person Games, Admissibility, Monotonicity, and Informedness ..................................................... 134 6.5 Heuristics and Complexity ....................................................................................................................................... 141 6.6 Summary ........................................................................................................................................................................ 145 6.7 Exercises ......................................................................................................................................................................... 145 III The Propositional and Predicate Calculi and Resolution- Based Reasoning 7 Introduction to the Propositional and Predicate Calculi ............................................................ 151 7.1 The Propositional Calculus ....................................................................................................................................... 152 7.2 The Predicate Calculus ............................................................................................................................................... 156 7.3 Summary ........................................................................................................................................................................ 169 7.4 Exercises ......................................................................................................................................................................... 169 8 The Predicate Calculus and Unification .................................................................................................. 171 8.1 Using Reasoning Rules to Produce New Predicate Expressions .................................................................. 172 8.2 The Unification Algorithm ........................................................................................................................................ 176 8.3 Summary ........................................................................................................................................................................ 189 8.4 Exercises ......................................................................................................................................................................... 189 9 Resolution: Reasoning with the Propositional and Predicate Calculi................................ 191 9.1 Introduction to Resolution ....................................................................................................................................... 192 9.2 Resolution Refutation Systems............................................................................................................................... 195 9.3 Strategies and Simplification Techniques for Resolution .............................................................................. 204 9.4 Extracting Answers from Resolution Refutations ............................................................................................ 209 9.5 Logic Programming and Prolog ............................................................................................................................. 212 9.6 Summary ........................................................................................................................................................................ 219 9.7 Exercises ......................................................................................................................................................................... 219 IV Advanced Applications of Symbol-Based AI 10 The Production System Representation and Search Engine .................................................... 223 10.1 The Production System: An Architecture for Organizing Search ................................................................ 224 10.2 More Examples of Production System Problem-Solving ............................................................................... 227 10.3 The Expert System ...................................................................................................................................................... 233 10.4 The Physical Symbol System Hypothesis and the Birth of Cognitive Science ........................................ 238 10.5 Summary ........................................................................................................................................................................ 240 10.6 Exercises ......................................................................................................................................................................... 241 XIV Contents

11 Advanced Applications of Symbol-Based AI: Planning and Learning ............................... 243 11.1 Introduction to Planning and Robotics ............................................................................................................... 244 11.2 Using Planning Macros: STRIPS .............................................................................................................................. 250 11.3 Model-Based Planning: A NASA Example (Williams and Nayak 1996, 1997) ........................................... 252 11.4 Symbol-Based Learning ............................................................................................................................................ 256 11.5 Bacon: Modeling the Celestial Environment ...................................................................................................... 268 11.6 Expertise Wherever It Is Needed ............................................................................................................................ 269 11.7 Summary ........................................................................................................................................................................ 271 11.8 Exercises ......................................................................................................................................................................... 271 12 Uncertain Reasoning: Symbol Based ....................................................................................................... 273 12.1 Logic-Based Reasoning in Uncertain Situations ............................................................................................... 274 12.2 Uncertain Reasoning: Alternatives to Logic ....................................................................................................... 286 12.3 Summary and Pointers to Parts VI and VII .......................................................................................................... 295 12.4 Exercises ......................................................................................................................................................................... 295 V Symbol-Based Associational Models for AI 13 Introduction to Association-Based Knowledge Representations ........................................ 299 13.1 The Behaviorist Tradition and Semantic Networks.......................................................................................... 300 13.2 Conceptual Dependencies ....................................................................................................................................... 307 13.3 Scripts .............................................................................................................................................................................. 313 13.4 Summary ........................................................................................................................................................................ 318 13.5 Exercises ......................................................................................................................................................................... 318 14 Association-Based Representations: Frames, Conceptual Graphs, WordNet, and FrameNet .................................................................................................................................................................... 321 14.1 Frames ............................................................................................................................................................................. 322 14.2 Conceptual Graphs ..................................................................................................................................................... 326 14.3 Current Uses of Association-Based Representations ...................................................................................... 336 14.4 Summary ........................................................................................................................................................................ 338 14.5 Exercises ......................................................................................................................................................................... 338 VI Neural or Connectionist Networks 15 An Introduction to Neural Networks ........................................................................................................ 345 15.1 An Artificial Neuron and Applications ................................................................................................................. 346 15.2 Early Research: McCulloch, Pitts, and Hebb ....................................................................................................... 348 15.3 Perceptrons ................................................................................................................................................................... 354 15.4 Summary ........................................................................................................................................................................ 360 15.5 Exercises ......................................................................................................................................................................... 361 16 The Delta Rule, Backpropagation, and Matrix Representations ........................................... 363 16.1 The Generalized Delta Rule ...................................................................................................................................... 364 16.2 The Backpropagation Algorithm ........................................................................................................................... 368 16.3 Matrix Representations for Network Processing .............................................................................................. 374 16.4 Matrix Representations and Neural Network Solutions ................................................................................ 377 16.5 Summary ........................................................................................................................................................................ 382 16.6 Exercises ......................................................................................................................................................................... 382 XV Contents

17 Deep Learning: Introduction and Representations ....................................................................... 383 17.1 Toward Deep Learning .............................................................................................................................................. 384 17.2 Meta-parameters for Very Large Networks ........................................................................................................ 392 17.3 Alternative Architectures: Convolutional and Recurrent Networks .......................................................... 397 17.4 Autoencoders and Transfer Learning ................................................................................................................... 403 17.5 Summary ........................................................................................................................................................................ 406 17.6 Exercises ......................................................................................................................................................................... 407 18 Building Language Models and Transformers ................................................................................... 409 18.1 Latent Semantic Analysis: Distributed Semantic Representations ............................................................ 410 18.2 Building Large Language Models .......................................................................................................................... 412 18.3 Toward Transformer-Based Large Language Models ..................................................................................... 417 18.4 The Transformer in Practice ..................................................................................................................................... 428 18.5 Summary ........................................................................................................................................................................ 437 18.6 Exercises ......................................................................................................................................................................... 438 19 Alternative Network Architectures: Prototypes and Classifiers ............................................ 441 19.1 A Kohonen Network: Winner-Takes-All Classification ..................................................................................... 442 19.2 A Kohonen Network: Learning Prototypes......................................................................................................... 444 19.3 Outstar Networks and Counterpropagation ...................................................................................................... 446 19.4 Supervised Hebbian Learning ................................................................................................................................ 450 19.5 Associative Memories and the Linear Associator ............................................................................................. 452 19.6 Summary ........................................................................................................................................................................ 457 19.7 Exercises ......................................................................................................................................................................... 457 20 Alternative Network Architectures: Attractor Networks and Memories ......................... 459 20.1 Introduction to Associative Memories ................................................................................................................. 460 20.2 BAM, the Bidirectional Associative Memory ...................................................................................................... 461 20.3 Autoassociative Memories and Hopfield Networks ........................................................................................ 467 20.4 Summary ........................................................................................................................................................................ 472 20.5 Exercises ......................................................................................................................................................................... 472 VII Probabilistic Artificial Intelligence 21 Counting, the Foundation for Probabilities ........................................................................................ 475 21.1 Introduction to Probabilistic Reasoning ............................................................................................................. 476 21.2 The Elements of Counting ........................................................................................................................................ 478 21.3 Elements of Probability Theory .............................................................................................................................. 482 21.4 Summary ........................................................................................................................................................................ 488 21.5 Exercises ......................................................................................................................................................................... 488 22 Bayes’ Theorem ....................................................................................................................................................... 491 22.1 Random Variables ....................................................................................................................................................... 492 22.2 Conditional Probability and an Introduction to Bayesian Reasoning ....................................................... 495 22.3 Bayes’ Theorem ............................................................................................................................................................ 499 22.4 Two Examples of Bayesian Reasoning ................................................................................................................. 503 22.5 Summary ........................................................................................................................................................................ 505 22.6 Exercises ......................................................................................................................................................................... 505 XVI Contents

23 Bayesian Belief Networks and Observable Markov Models ..................................................... 507 23.1 Introduction to Stochastic Models ........................................................................................................................ 508 23.2 A Directed Graphical Model: The Bayesian Belief Network .......................................................................... 509 23.3 Dynamic Bayesian Networks ................................................................................................................................... 517 23.4 Observable Markov Models ..................................................................................................................................... 518 23.5 Summary ........................................................................................................................................................................ 522 23.6 Exercises ......................................................................................................................................................................... 522 24 Hidden Markov and Alternative Probabilistic Models ................................................................. 525 24.1 Hidden Markov Models ............................................................................................................................................. 526 24.2 Important Variants of Hidden Markov Models .................................................................................................. 529 24.3 A Short Survey of Alternative Markov Models .................................................................................................. 534 24.4 First-Order Alternatives to BBNs and HMMs ...................................................................................................... 536 24.5 Two Stochastic Engineering Examples: An MDP and a DBN ........................................................................ 539 24.6 Summary ........................................................................................................................................................................ 544 24.7 Exercises ......................................................................................................................................................................... 545 VIII AI: Ethical Issues, Fundamental Limitations, and Future Promise 25 Artificial Intelligence: User’s Ethical Issues .......................................................................................... 551 25.1 Artificial and Human Intelligence .......................................................................................................................... 552 25.2 Ethical Issues from the Perspective of the AI User ........................................................................................... 556 25.3 We Are Already Embedded in an AI World.......................................................................................................... 563 25.4 Summary ........................................................................................................................................................................ 570 25.5 Exercises ......................................................................................................................................................................... 571 26 AI Ethical Issues: From a Social Perspective ........................................................................................ 573 26.1 Building Society-Oriented AI Projects .................................................................................................................. 574 26.2 If We Do Build These AI Projects, What Levels of Protection Are Needed? .............................................. 577 26.3 What Is Data? ................................................................................................................................................................ 581 26.4 Data and Algorithm Transparency and Explainability.................................................................................... 582 26.5 Design Honesty, Transparency, Explanations, Data Bias, and Privacy ...................................................... 592 26.6 Safeguarding AI Research and Practice ............................................................................................................... 595 26.7 Summary ........................................................................................................................................................................ 597 26.8 Exercises ......................................................................................................................................................................... 598 27 AI: Philosophical Perspectives, Current Limitations, and Future Promise ..................... 601 27.1 Modern AI: A Psychological, Mathematical, and Philosophical Perspective .......................................... 602 27.2 Several Fundamental Limitations of Current AI Technology ........................................................................ 607 27.3 Artificial Intelligence: Where Are We Going? ..................................................................................................... 610 Supplementary Information Bibliography ................................................................................................................................................................ 616 Index ................................................................................................................................................................................ 633 XVII Contents

I Introducing Artificial Intelligence We begin with an historical introduction to artificial intelligence. The roots of our technology can be found in the philosophical, mathematical, and engineering traditions of Western culture. 7 Chapter 1 opens with the musings of Mary Shelly, the author of Frankenstein or The Modern Prometheus. This gothic novel not only offers a view of the modern project of creating intelligent artifacts, but also connects us back to the Promethean myth. Prometheus was tortured by the gods and tied forever to a boulder. His crime was bringing the secrets of the gods to humankind, including the knowledge of science, medicine, and the mineral wealth of the natural world. The AI challenge is to deliver to society this same knowledge and the skills for its use! 7 Chapter 1 introduces the philosophical positions and many of the mathematical traditions supporting modern artificial intelli- gence. These include the early writings of George Boole, Ada Lovelace, and the ideas of the Logical Positivists and American Pragmatists. The chapter also describes Alan Turing’s test for deter- mining whether a computer can be seen as intelligent. 7 Chapter 1 concludes with a projection for AI’s future, described by the tasks proposed for the Dartmouth Summer Workshop of 1956. 7 Chapter 2 begins with the task of finding a definition for the discipline of artificial intelligence. We next discuss the AI project of trying to represent important features of the natural world for a computational device. Finally, we describe the major representa- tional approaches AI has taken over the 70 years since its incep- tion. These are the symbol- based, neural network, stochastic, and genetic/emergent representations of the natural world. The remainder of our book fills out the details of this story. 1

Many of the publications supporting the foundations of comput- ing and artificial intelligence are available on the internet and accessible for the interested reader. If there is to be a program- ming component supporting the presentation of later parts of this text, it is important at this early stage of the AI story to make these materials available to those wishing to use them. Contents Chapter 1 The Pre-History of Artificial Intelligence – 3 Chapter 2 Computing, Representations, and Definitions of Artificial Intelligence – 27

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2025 G. F. Luger, Artificial Intelligence: Principles and Practice, https://doi.org/10.1007/978-3-031-57437-5_1 The Pre-History of Artificial Intelligence Contents 1.1 Mary Shelley, Frankenstein, and Prometheus – 4 1.2 The Age of Rationalism – 6 1.3 The Empiricist Tradition – 8 1.4 Immanuel Kant: Bridging the Rationalist/Empiri- cist Viewpoints – 9 1.5 The Reverend Thomas Bayes – 11 1.6 The Mathematical Foundations for Artificial Intel- ligence – 11 1.7 American Pragmatism – 16 1.8 The Turing Test: Can a Machine Be “Intelligent”? – 17 1.9 The 1956 Dartmouth Summer Workshop – 20 1.10 Summary – 24 1.11 Exercises – 24 13

1 Hear the rest, and you will marvel even more at the crafts and resources I have contrived. Greatest was this: in the former times if a man fell sick he had no defense against the sickness, neither healing food nor drink, nor unguent; but through the lack of drugs men wasted away, until I showed them the blending of mild simples wherewith they drive out all manner of diseases . . . . It was I who made visible to men’s eyes the flaming signs of the sky that were before dim. So much for these. Beneath the earth, man’s hidden blessing, copper, iron, silver, and gold—will anyone claim to have discovered these before I did? No one, I am very sure, who wants to speak truly and to the purpose. One brief word will tell the whole story: all arts that mortals have come from Prometheus. AESCHYLUS, Prometheus Bound All people by nature desire to know... ARISTOTLE, Opening sentence of the Metaphysics 7 Chapter 1 offers a brief history of the ideas that support and enable the discipline of artificial intelligence. First, we describe the mythology behind the use of knowledge and skill to improve the human condition. This claimed usurpation of the powers of the gods by Aeschylus, see chapter heading, is often seen as a challenge to common sense as well as an act of human arrogance. Second, we describe the seventeenth through twen- tieth century mathematics and engineering that laid the foun- dation for the sciences of computing and artificial intelligence. We then present Alan Turing’s test for intelligence published in 1950 in the journal Mind. Finally, we describe the 1956 Dartmouth Summer Workshop where the name “artificial intel- ligence” was adopted and where, for the first time, AI research areas were identified and described. 1.1 Mary Shelley, Frankenstein, and Prometheus The Greek dramatist Aeschylus speaks, through the words of Prometheus, whose name in Greek means “fore knowing,” of the benefits mankind has received through his “transgressions.” Prometheus’s purpose was not only to steal fire from the gods for use by the human race but to enlighten humanity through knowledge or nous, the rational mind. This use of knowledge or intelligence forms the foundation for all of human technology as well as offers promise and hope for human civilization. The plays of Aeschylus illustrate a deep awareness of the extraordinary power of knowledge. The benefits of artificial intelligence over the past decades have been applied to all the 4 Chapter 1 · The Pre-History of Artificial Intelligence