ebook img

Introduction to Artificial Intelligence PDF

365 Pages·2018·5.791 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Introduction to Artificial Intelligence

Undergraduate Topics in Computer Science Wolfgang Ertel Introduction to Artificial Intelligence Second Edition Undergraduate Topics in Computer Science Series editor Ian Mackie Advisory Board SamsonAbramsky, University ofOxford, Oxford,UK KarinBreitman, PontificalCatholic University of RiodeJaneiro,RiodeJaneiro, Brazil Chris Hankin,Imperial CollegeLondon,London,UK Dexter Kozen,Cornell University, Ithaca, USA AndrewPitts, University ofCambridge, Cambridge, UK Hanne Riis Nielson,Technical University of Denmark,Kongens Lyngby,Denmark StevenSkiena, StonyBrook University,StonyBrook, USA Iain Stewart, University of Durham, Durham, UK UndergraduateTopics inComputerScience (UTiCS) delivers high-qualityinstruc- tionalcontentforundergraduatesstudyinginallareasofcomputingandinformation science. From core foundational and theoretical material to final-year topics and applications,UTiCSbookstakeafresh,concise,andmodernapproachandareideal for self-study or for a one- or two-semester course. The texts are all authored by establishedexpertsintheirfields,reviewedbyaninternationaladvisoryboard,and containnumerousexamplesandproblems.Manyincludefullyworkedsolutions. More information about this series at http://www.springer.com/series/7592 Wolfgang Ertel Introduction to Artificial Intelligence Second Edition Translated by Nathanael Black With illustrations by Florian Mast 123 WolfgangErtel Hochschule Ravensburg-Weingarten Weingarten Germany ISSN 1863-7310 ISSN 2197-1781 (electronic) Undergraduate Topics inComputer Science ISBN978-3-319-58486-7 ISBN978-3-319-58487-4 (eBook) DOI 10.1007/978-3-319-58487-4 LibraryofCongressControlNumber:2017943187 1stedition:©Springer-VerlagLondonLimited2011 2ndedition:©SpringerInternationalPublishingAG2017 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar methodologynowknownorhereafterdeveloped. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse. 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 authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations. Printedonacid-freepaper ThisSpringerimprintispublishedbySpringerNature TheregisteredcompanyisSpringerInternationalPublishingAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface to the Second Edition After 60 years, Artificial Intelligence (AI) has now reached industry and the con- sciousness of the population. The impressive successes and new AI methods are nowsorelevantthattheyshouldbetaughteveninabasiccourse.Inabout30new pages,Ireportmainlyondeeplearning,aconsistentfurtherdevelopmentofneural networks,whichfinallyenablesimageprocessing systemstorecognizealmostany object in pixel images. Among other benefits, this lead to the first computer pro- gram that could beat one of the world’s best Go players. InthenewsectiononDeepLearning,wemustnotleaveoutashortreportabout the fascinating new subarea of creativity. For the first time neural networks can creatively generate texts, music pieces, and even paintings in the style of the old masters. These achievements are based on many years of research on neural net- works and machine learning. Practical AI has developed into an engineering dis- cipline in which programs are developed in large industrial teams by experts from various specializations. Self-drivingcars,servicerobots,andsmarthomes—whichareallapplicationsof AI—willgreatlychangeourlives.However,inadditiontogreatraysofhope,there will be a dark side. Though we live in a time of rapid technological progress, we have long since exceeded the limits of growth. We must therefore think about sustainability when implementing each new invention. In Chap. 1, I would like to give you some food for thought about this topic. Othernewadditionstothebookincludeasectiononperformanceevaluationof clusteringalgorithmsandtwopracticalexamplesexplainingBayes’theoremandits relevanceineverydaylife.Finally,inasectiononsearchalgorithms,weanalyzethe cycle check, explain route planning for car navigation systems, and briefly intro- duce Monte Carlo Tree Search. All known errors have been corrected and updates have been made in many places. Iwouldlike tosincerely thank thereaders whohavegivenmefeedback andall those who contributed to this new edition through proofreading and suggestions. v vi PrefacetotheSecondEdition IwouldespeciallyliketothankAdrianBatzillfortherouteplanningmeasurements and graphs, as well as Nate Black, Nicole Dathe, Markus Schneider, Robin Leh- mann, Ankita Agrawal, Wenzel Massag, Lars Berge, Jonas Lang, and Richard Cubek. Ravensburg Wolfgang Ertel March 2017 Preface to the First Edition Artificial Intelligence (AI) has the definite goal of understanding intelligence and building intelligent systems. However, the methods and formalisms used on the way to this goal are not firmly set, which has resulted in AI consisting of a multitudeofsubdisciplinestoday.ThedifficultyinanintroductoryAIcourseliesin conveying as many branches as possible without losing too much depth and precision. RussellandNorvig’sbook[RN10]ismoreorlessthestandardintroductioninto AI. However, since this book has 1,152 pages, and since it is too extensive and costlyformoststudents,therequirementsforwritingthisbookwereclear:itshould be an accessible introduction tomodern AI for self-study orasthe foundationof a four-hour lecture, with at most 300 pages. The result is in front of you. Inthespaceof300pages,afieldasextensiveasAIcannotbefullycovered.To avoid turning the book into a table of contents, I have attempted to go into some depthandtointroduceconcretealgorithmsandapplicationsineachofthefollowing branches: agents, logic, search, reasoning with uncertainty, machine learning, and neural networks. Thefieldsofimageprocessing,fuzzylogic,andnaturallanguageprocessingare not covered in detail. The field of image processing, which is important for all of computer science, is a stand-alone discipline with very good textbooks, such as [GW08]. Natural language processing has a similar status. In recognizing and generating text and spoken language, methods from logic, probabilistic reasoning, and neural networks are applied. In this sense this field is part of AI. On the other hand,computerlinguisticsisitsownextensivebranchofcomputerscienceandhas much in common with formal languages. In this book we will point to such appropriatesystemsinseveralplaces,butnotgive asystematic introduction.Fora firstintroductioninthisfield,werefertoChaps.22and23in[RN10].Fuzzylogic, orfuzzysettheory,hasdevelopedintoabranchofcontroltheoryduetoitsprimary application in automation technology and is covered in the corresponding books and lectures. Therefore we will forego an introduction here. The dependencies between chapters of the book are coarsely sketched in the graphshownbelow.Tokeepitsimple,Chap.1,withthefundamentalintroduction for all further chapters, is left out. As an example, the thicker arrow from 2 to 3 means that propositional logic is a prerequisite for understanding predicate logic. vii viii PrefacetotheFirstEdition The thin arrow from 9 to 10 means that neural networks are helpful for under- standing reinforcement learning, but not absolutely necessary. Thin backward arrows shouldmake clear thatlaterchapterscangive more depth ofunderstanding to topics which have already been learned. Thisbookisapplicabletostudentsofcomputerscienceandothertechnicalnatural sciences and, for the most part, requires high school level knowledge of mathe- matics. In several places, knowledge from linear algebra and multidimensional analysisisneeded.Foradeeperunderstandingofthecontents,activelyworkingon the exercises is indispensable. This means that the solutions should only be con- sulted after intensive work with each problem, and only to check one’s solutions, true to Leonardo da Vinci’s motto “Study without devotion damages the brain”. Somewhat more difficult problems are marked with ❄, and especially difficult ones with ❄❄. Problems which require programming or special computer science knowledge are labeled with ➳. On the book’s web site at http://www.hs-weingarten.de/*ertel/aibook digital materialsfortheexercisessuchastrainingdataforlearningalgorithms,apagewith references to AI programs mentioned in the book, a list of links to the covered topics,aclickablelistofthebibliography,anerratalist,andpresentationslidesfor lecturers can be found. I ask the reader to please send suggestions, criticisms, and tips about errors directly to [email protected]. ThisbookisanupdatedtranslationofmyGermanbook“GrundkursKünstliche Intelligenz” published by Vieweg Verlag. My special thanks go to the translator NathanBlackwhoinanexcellenttrans-AtlanticcooperationbetweenGermanyand California via SVN, Skype and Email produced this text. I am grateful to Franz Kurfeß, who introduced me to Nathan; to MatthewWight for proofreading the translated book and to Simon Rees from Springer Verlag for his patience. I would like to thank my wife Evelyn for her support and patience during this time consuming project. Special thanks go to Wolfgang Bibel and Chris Loben- schuss, who carefully corrected the German manuscript. Their suggestions and discussions lead to many improvements and additions. For reading the corrections and other valuable services, I would like to thank Richard Cubek, Celal Döven, Joachim Feßler, Nico Hochgeschwender, Paul Kirner, Wilfried Meister, Norbert Perk,PeterRadtke,MarkusSchneider,ManfredSchramm,UliStärk,MichelTokic, Arne Usadel andall interestedstudents. Mythanks also goouttoFlorian Mastfor the priceless cartoons and very effective collaboration. Ihopethatduringyourstudiesthisbookwillhelpyousharemyfascinationwith Artificial Intelligence. Ravensburg Wolfgang Ertel February 2011 Contents 1 Introduction... .... .... ..... .... .... .... .... .... ..... .... 1 1.1 What Is Artificial Intelligence?.... .... .... .... ..... .... 1 1.1.1 Brain Science and Problem Solving.. .... ..... .... 3 1.1.2 The Turing Test and Chatterbots.... .... ..... .... 5 1.2 The History of AI. ..... .... .... .... .... .... ..... .... 5 1.2.1 The First Beginnings. .... .... .... .... ..... .... 7 1.2.2 Logic Solves (Almost) All Problems. .... ..... .... 8 1.2.3 The New Connectionism.. .... .... .... ..... .... 9 1.2.4 Reasoning Under Uncertainty .. .... .... ..... .... 9 1.2.5 Distributed, Autonomous and Learning Agents.. .... 10 1.2.6 AI Grows Up... .... .... .... .... .... ..... .... 11 1.2.7 The AI Revolution... .... .... .... .... ..... .... 11 1.3 AI and Society ... ..... .... .... .... .... .... ..... .... 11 1.3.1 Does AI Destroy Jobs? ... .... .... .... ..... .... 11 1.3.2 AI and Transportation .... .... .... .... ..... .... 14 1.3.3 Service Robotics .... .... .... .... .... ..... .... 15 1.4 Agents . .... .... ..... .... .... .... .... .... ..... .... 17 1.5 Knowledge-Based Systems... .... .... .... .... ..... .... 19 1.6 Exercises.... .... ..... .... .... .... .... .... ..... .... 20 2 Propositional Logic. .... ..... .... .... .... .... .... ..... .... 23 2.1 Syntax.. .... .... ..... .... .... .... .... .... ..... .... 23 2.2 Semantics ... .... ..... .... .... .... .... .... ..... .... 24 2.3 Proof Systems.... ..... .... .... .... .... .... ..... .... 26 2.4 Resolution... .... ..... .... .... .... .... .... ..... .... 30 2.5 Horn Clauses .... ..... .... .... .... .... .... ..... .... 33 2.6 Computability and Complexity.... .... .... .... ..... .... 36 2.7 Applications and Limitations . .... .... .... .... ..... .... 37 2.8 Exercises.... .... ..... .... .... .... .... .... ..... .... 37 ix

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.