ebook img

Theoretical Foundations of Artificial General Intelligence PDF

331 Pages·2012·4.478 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 Theoretical Foundations of Artificial General Intelligence

ATLANTIS THINKINGMACHINES VOLUME4 SERIESEDITOR: KAI-UWEKÜHNBERGER Atlantis Thinking Machines SeriesEditor: Kai-UweKühnberger InstituteofCognitiveScience UniversityofOsnabrück,Germany (ISSN:1877-3273) Aimsandscopeoftheseries This series publishes books resulting from theoretical research on and reproductions of generalArtificial Intelligence(AI). Thebookseries focusesonthe establishmentof new theoriesandparadigmsinAI.Atthesametime,theseriesaimsatexploringmultiplescien- tificanglesandmethodologies,includingresultsfromresearchincognitivescience,neu- roscience,theoreticalandexperimentalAI,biologyandfrominnovativeinterdisciplinary methodologies. Formoreinformationonthisseriesandourotherbookseries,pleasevisitourwebsiteat: www.atlantis-press.com/publications/books AMSTERDAM–PARIS –BEIJING (cid:2)c ATLANTISPRESS Theoretical Foundations of Artificial General Intelligence PeiWang(Ed.) DepartmentofComputerandInformationSciences,TempleUniversity, 1805N.BroadStreet,Philadelphia,PA19122,USA BenGoertzel(Ed.) NovamenteLLC/BiomindLLC, 1405BernerdPlace,Rockville,MD20851,USA AMSTERDAM–PARIS –BEIJING AtlantisPress 8,squaredesBouleaux 75019Paris,France ForinformationonallAtlantisPresspublications,visitourwebsiteat:www.atlantis-press.com Copyright Thisbook,oranypartsthereof,maynotbereproducedforcommercialpurposesinanyformorby anymeans,electronicormechanical,includingphotocopying,recordingoranyinformationstorage andretrievalsystemknownortobeinvented,withoutpriorpermissionfromthePublisher. AtlantisThinkingMachines Volume1:Enaction,Embodiment,EvolutionaryRobotics.SimulationModelsforaPost-Cognitivist ScienceofMind-MariekeRohde,EzequielA.DiPaolo Volume2:Real-WorldReasoning:TowardScalable,UncertainSpatiotemporal,Contextualand CausalInference-BenGoertzel,NilGeisweiller,LúcioCoelho,PredragJanicic,CassioPennachin Volume3:IntegrationofWorldKnowledgeforNaturalLanguageUnderstanding-Ekaterina Ovchinnikova ISBNs Print: 978-94-91216-61-9 E-Book: 978-94-91216-62-6 ISSN: 1877-3273 (cid:2)c 2012ATLANTISPRESS Contents 1. Introduction 1 PeiWangandBenGoertzel 1.1 TheMatterofArtificialGeneralIntelligence . . . . . . . . . . . . . . . . . . . . . 1 1.2 TheMatterofTheoreticalFoundation . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3 TheMatterofObjective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.4 TheMatterofApproach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.5 ChallengesattheHeartoftheMatter . . . . . . . . . . . . . . . . . . . . . . . . 6 1.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2. ArtificialIntelligenceandCognitiveModelingHavetheSameProblem 11 NicholasL.Cassimatis 2.1 TheIntelligenceProblem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2 ExistingMethodsandStandardsarenotSufficient. . . . . . . . . . . . . . . . . . 13 2.3 CognitiveModeling:TheModelFitImperative . . . . . . . . . . . . . . . . . . . 16 2.4 ArtificialIntelligenceandCognitiveModelingCanHelpEachOther . . . . . . . . 21 2.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3. ThePiaget-MacGuyverRoom 25 SelmerBringsjordandJohnLicato 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.2 MoreonPsychometricAGI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.3 Descartes’TwoTests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.4 Piaget’sViewofThinking&TheMagnetTest . . . . . . . . . . . . . . . . . . . . 35 3.5 TheLISAmodel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.6 Analogico-DeductiveReasoningintheMagnetTest . . . . . . . . . . . . . . . . . 39 3.7 NextSteps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 v vi TheoreticalFoundationsofArtificialGeneralIntelligence 4. BeyondtheOctopus:FromGeneralIntelligencetowardaHuman-likeMind 49 SamS.AdamsandSteveBurbeck 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.2 OctopusIntelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 4.3 A“Ladder”ofIntelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.4 LinguisticGrounding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 4.5 ImplicationsoftheLadderforAGI . . . . . . . . . . . . . . . . . . . . . . . . . . 57 4.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 5. OneDecadeofUniversalArtificialIntelligence 67 MarcusHutter 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 5.2 TheAGIProblem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5.3 UniversalArtificialIntelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 5.4 FacetsofIntelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 5.5 SocialQuestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 5.6 StateoftheArt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 5.7 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 6. DeepReinforcementLearningasFoundationforArtificialGeneralIntelligence 89 ItamarArel 6.1 Introduction:DecomposingtheAGIProblem . . . . . . . . . . . . . . . . . . . . 89 6.2 DeepLearningArchitectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 6.3 ScalingDecisionMakingunderUncertainty . . . . . . . . . . . . . . . . . . . . . 94 6.4 NeuromorphicDevicesScalingAGI . . . . . . . . . . . . . . . . . . . . . . . . . 99 6.5 ConclusionsandOutlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 7. TheLIDAModelasaFoundationalArchitectureforAGI 103 UsefFaghihiandStanFranklin 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 7.2 WhytheLIDAModelMayBeSuitableforAGI . . . . . . . . . . . . . . . . . . . 104 7.3 LIDAArchitecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 7.4 CognitiveArchitectures,FeaturesandtheLIDAModel . . . . . . . . . . . . . . . 108 7.5 Discussion,Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Contents vii 8. TheArchitectureofHuman-LikeGeneralIntelligence 123 BenGoertzel,M.Iklé,andJ.Wigmore 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 8.2 KeyIngredientsoftheIntegrativeHuman-LikeCognitiveArchitectureDiagram . . 125 8.3 AnArchitectureDiagramforHuman-LikeGeneralIntelligence. . . . . . . . . . . 128 8.4 InterpretationandApplicationoftheIntegrativeDiagram . . . . . . . . . . . . . . 136 8.5 CognitiveSynergy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 8.6 WhyIsItSoHardtoMeasurePartialProgressTowardHuman-LevelAGI? . . . . 140 8.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 9. ANewConstructivistAI 145 KristinnR.Thórisson 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 9.2 TheNatureof(General)Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . 147 9.3 ConstructionistAI:ACriticalLook . . . . . . . . . . . . . . . . . . . . . . . . . 151 9.4 TheCallforaNewMethodology . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 9.5 TowardsaNewConstructivistAI. . . . . . . . . . . . . . . . . . . . . . . . . . . 158 9.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 10. TowardsanActualGödelMachineImplementation 173 BasR.SteunebrinkandJürgenSchmidhuber 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 10.2 TheGödelMachineConcept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 10.3 TheTheoreticalFoundationsofSelf-ReflectiveSystems . . . . . . . . . . . . . . 178 10.4 NestedMeta-CircularEvaluators . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 10.5 AFunctionalSelf-ReflectiveSystem . . . . . . . . . . . . . . . . . . . . . . . . . 186 10.6 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Appendix:DetailsofNotationUsed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 11. ArtificialGeneralIntelligenceBeginswithRecognition 197 TsviAchler 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 11.2 EvaluatingFlexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 11.3 EvaluationofFlexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 11.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216 viii TheoreticalFoundationsofArtificialGeneralIntelligence 12. TheoryBlendingasaFrameworkforCreativityinSystemsforGeneralIntelligence 219 MaricarmenMartinezetal. 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 12.2 ProductivityandCognitiveMechanisms . . . . . . . . . . . . . . . . . . . . . . . 221 12.3 Cross-DomainReasoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 12.4 BasicFoundationsofTheoryBlending . . . . . . . . . . . . . . . . . . . . . . . . 226 12.5 TheComplexPlane:AChallengingHistoricalExample. . . . . . . . . . . . . . . 228 12.6 OutlookforNextGenerationGeneralIntelligentSystems . . . . . . . . . . . . . . 233 12.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 13. ModelingEmotionandAffect 241 JoschaBach 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 13.2 EmotionandAffect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244 13.3 AffectiveStatesEmergingfromCognitiveModulation . . . . . . . . . . . . . . . 246 13.4 Higher-LevelEmotionsEmergingfromDirectingValencedAffects. . . . . . . . . 250 13.5 GeneratingRelevance:theMotivationalSystem . . . . . . . . . . . . . . . . . . . 252 13.6 MotiveSelection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 13.7 PuttingitAllTogether . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 14. AGIandMachineConsciousness 263 AntonioChellaandRiccardoManzotti 14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 14.2 Consciousness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 14.3 MachineConsciousness. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267 14.4 Agent’sBody . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 14.5 InteractionswiththeEnvironment . . . . . . . . . . . . . . . . . . . . . . . . . . 271 14.6 Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 14.7 FreeWill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274 14.8 Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 14.9 Creativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 14.10 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 15. HumanandMachineConsciousnessasaBoundaryEffectintheConcept AnalysisMechanism 283 RichardLoosemore 15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 15.2 TheNatureofExplanation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 15.3 TheRealMeaningofMeaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 15.4 SomeFalsifiablePredictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301 15.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304 Contents ix 16. TheoriesofArtificialIntelligence 305 PeiWang 16.1 TheProblemofAITheory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305 16.2 NatureandContentofAITheories . . . . . . . . . . . . . . . . . . . . . . . . . . 308 16.3 DesiredPropertiesofaTheory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312 16.4 RelationsamongtheProperties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 16.5 IssuesontheProperties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 16.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 Index 325 Chapter1 Introduction: What Is the Matter Here? PeiWang1andBenGoertzel2 1TempleUniversity,USA 2NovamenteLLC,USA [email protected], [email protected] Thischapterprovidesageneralintroductiontothevolume,givinganoverviewoftheAGI fieldandthecurrentneedforexplorationandclarificationofitsfoundations, andbriefly summarizingthecontentsofthevariouschapters. 1.1 TheMatterofArtificialGeneralIntelligence ArtificialGeneralIntelligence(AGI),roughlyspeaking,referstoAIresearchanddevel- opmentinwhich“intelligence”isunderstoodasageneral-purposecapability,notrestricted toanynarrowcollectionofproblemsordomains,andincludingtheabilitytobroadlygen- eralizetofundamentallynewareas[4]. TheprecisedefinitionofAGIispartofthesubject matteroftheAGIfield,anddifferenttheoreticalapproachestoAGImayembodydifferent slants on the very conceptof AGI. In practicalterms, though, the variousresearchersin thefieldshareastrongcommonintuitionregardingthecoreconcernsofAGI–andhowit differsfromthe“narrowAI”thatcurrentlydominatestheAIfield. In the earliest daysof AI research, in the middle of the last century, the objective of thefieldwastobuild“thinkingmachines”withcapacitycomparabletothatofthehuman mind[2,6,9].InthedecadesfollowingthefoundingoftheAIfield,variousattemptsarose to attack the problem of human-levelartificial general intelligence, such as the General ProblemSolver[7]andtheFifthGenerationComputerSystems[3]. Theseearlyattempts failedtoreachtheiroriginalgoals,andintheviewofmostAIresearchers,failedtomake dramaticconceptualorpracticalprogresstowardtheirgoals. Basedontheseexperiences, 1

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.