Mental Models and Their Dynamics, Adaptation, and Control: A Self-Modeling Network Modeling Approach PDF
Preview Mental Models and Their Dynamics, Adaptation, and Control: A Self-Modeling Network Modeling Approach
Studies in Systems, Decision and Control 394 Jan Treur Laila Van Ments Editors Mental Models and Their Dynamics, Adaptation, and Control A Self-Modeling Network Modeling Approach Studies in Systems, Decision and Control Volume 394 SeriesEditor JanuszKacprzyk,SystemsResearchInstitute,PolishAcademyofSciences, Warsaw,Poland The series “Studies in Systems, Decision and Control” (SSDC) covers both new developments and advances, as well as the state of the art, in the various areas of broadly perceived systems, decision making and control–quickly, up to date and with a high quality. The intent is to cover the theory, applications, and perspec- tives on the state of the art and future developments relevant to systems, decision making, control, complex processes and related areas, as embedded in the fields of engineering, computer science, physics, economics, social and life sciences, as wellastheparadigmsandmethodologiesbehindthem.Theseriescontainsmono- graphs, textbooks, lecture notes and edited volumes in systems, decision making and control spanning the areas of Cyber-Physical Systems, Autonomous Systems, SensorNetworks,ControlSystems,EnergySystems,AutomotiveSystems,Biolog- ical Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems, Robotics,SocialSystems,EconomicSystemsandother.Ofparticularvaluetoboth thecontributorsandthereadershiparetheshortpublicationtimeframeandtheworld- widedistributionandexposurewhichenablebothawideandrapiddisseminationof researchoutput. IndexedbySCOPUS,DBLP,WTIFrankfurteG,zbMATH,SCImago. AllbookspublishedintheseriesaresubmittedforconsiderationinWebofScience. Moreinformationaboutthisseriesathttps://link.springer.com/bookseries/13304 · Jan Treur Laila Van Ments Editors Mental Models and Their Dynamics, Adaptation, and Control A Self-Modeling Network Modeling Approach Editors JanTreur LailaVanMents SocialAIgroup AutoLeadStar,Jerusalem,Israel DepartmentofComputerScience VrijeUniversiteitAmsterdam,Amsterdam Noord-Holland,TheNetherlands ISSN2198-4182 ISSN2198-4190 (electronic) StudiesinSystems,DecisionandControl ISBN978-3-030-85820-9 ISBN978-3-030-85821-6 (eBook) https://doi.org/10.1007/978-3-030-85821-6 ©TheEditor(s)(ifapplicable)andTheAuthor(s),underexclusivelicensetoSpringerNature SwitzerlandAG2022 Thisworkissubjecttocopyright.AllrightsaresolelyandexclusivelylicensedbythePublisher,whether thewholeorpartofthematerialisconcerned,specificallytherightsoftranslation,reprinting,reuse ofillustrations,recitation,broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,and transmissionorinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilar ordissimilarmethodologynowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressedorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional claimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface Formalisinghumanbehaviourandmentalprocessesinaformalisedformatthatcan beusedbycomputationaldevicesisoneoftheholygrailsinthefieldsofartificial intelligenceandcognitivescience.Thehumanbrainissodynamicandcomplexthatit ishardtocomprehendthatacomputercanmimichowitworks,anideathathasbeen thebasisformanyfantasies,booksandmovies.Thisbookismeantasacontribution tothejourneytothisholygrailandtomakingthegapbetweenhumanbehaviourand computationaldevicesabitsmaller,withafocusinparticularonmentalprocesses involvingmentalmodels. Our interests in mental models and their internal simulation and learning were inspiredbyanumberofsources,includingliteratureinareassuchas (cid:129) Theoriesandpracticesinlearningpsychology,educationalscienceanddidactics andinparticularthelearningandadaptationofmentalmodelsaddressedinthis area (cid:129) Theconceptofas-ifbodyloopforinternalmentalsimulationofbodystatesfor emotions (cid:129) Theconceptofsimulatedperception–actionchainsforimaginationandconscious thought (cid:129) Theuseoftheoryofmindbasedoninternalmentalsimulation. These areas that all concern crucial elements of human mental processes and behaviourprovethatitisachallengetodesignformalisedandcomputationalmodels forsuchphenomenainvolvingmentalmodelsaccordingtosomegeneralarchitecture. Mentalmodelsareoftenconsideredasconsistingof(networksof)relationsbetween nodes. Such an architecture needs different types of representation for the states involved as nodes in a mental model (internal simulation can take place by their activation)butalsofortherelationsbetweenthesestates(thattriggertheactivationsof thestates).Whenmentalmodelsareusedforinternalsimulation,theirstateschange overtimebasedontheserelations.Moreover,theserelationsadaptwhenthemental modelislearnt,forgottenorrevisedovertime.Thissuggeststhattoaddressmental processes involving mental models, dynamic representations are required both for thestatesandfortherelations.Moreover,ontopofthatsomeformofmetacognition (metacognitivecontrol)isneededtocontroltheseinternalsimulationandadaptation v vi Preface processes. Together, these are non-trivial and challenging requirements to achieve anappropriatetypeofformalisedcomputationalarchitecture. In addition to the above-mentioned areas in the literature, our shared interest in mental models was influenced by a number of case studies addressing specific phenomena that involve mental models, for example, the influence of cognitive metaphors on behaviour and the influence of mental God models on behaviour. Moreover,theideathatapost-traumaticstressdisorderisbasedonalearntbutunde- sired mental model gave another inspiration. It is in these application cases that somefirstideasweredevelopedonhowcomputationalmodelscanbedesignedfor mentalprocessesinvolvingmentalmodels.Asafirststep,usinganetwork-oriented modellingapproach,examplenetworkmodelsweredevelopedforthesethreecases, where: (cid:129) A mental model was modelled by a subnetwork of an overall mental network model;thissubnetworkrepresentsthenodesandrelationsofthementalmodel. (cid:129) The use of a mental model for internal simulation as part of the overall mental processwasmodelledaspropagationofactivationwithinthissubnetworkrepre- sentingthementalmodel;outcomesofsuchformsofinternalsimulationthencan beusedasabasistodecideaboutactionsorbehaviour. (cid:129) Learning or adaptation of the mental model was modelled and implemented separatelyfromthenetworkmodel,tomakethenetworkanadaptivenetwork. (cid:129) Nometacognitivecontroloftheseprocesseswasexplicitlymodelled. Asasecondstep,itwasfoundouthowadaptivenetworkscanalsobemodelledin anetwork-orientedmannerbyusingso-calledself-modellingnetworks(alsocalled reified networks). Self-modelling networks include self-models with nodes repre- senting part of the network’s own structure, for example, nodes representing the weights of some of its connections. Using this notion, to introduce new adapta- tionprinciples,nonewimplementationshavetobemade,asthosenewadaptation principlescanbeaddressedatthemodellinglevelinstead.Byusingaself-modelling network,foramentalmodelwithitsnodesandrelationsmodelledasa(base)subnet- work,therelationsofthementalmodelcanexplicitlyberepresented(andadapted) byself-modelnodes.Thisindeedprovidesthedifferentrepresentationsneededfor mentalprocessesbasedonamentalmodel.Then,inadedicatedsoftwareenviron- ment,oneandthesamecomputationalself-modellingnetworkenginecanbeused in a generic manner to process both the internal simulation of a mental model as propagation of activation within the subnetwork for mental model and the mental modeladaptationaspropagationofactivationwithintheself-modelofthesubnet- workforthementalmodel.Moreover,astheconstructionofself-modellingnetworks can easily be iterated to obtain multiple orders of self-models, it was found that also metacognitive control can easily be modelled in this way by a second-order self-modelintheself-modelingnetwork. Inrecentyears,thisapproachhassuccessfullybeenappliedtomodeladiversityof specificmentalprocessesbasedonmentalmodels,asillustratedinmanychaptersin thisvolume.Inparticular,inPartsII(forindividualcases)andIII(forsocialcases), each including six chapters. These cover the three examples mentioned above in Preface vii Chap.5(PTSDasundesiredmentalmodel),Chap.10(cognitivemetaphorsasmental models)andChap.11(mentalGodmodels).OtherchaptersinPartIIaddressmental modelsforlearninghowacarworksandhowtodriveit(Chap.3),usingmetacog- nitivecontroltohandlemultiplementalmodels(Chap.4),learningbycounterfac- tualthinking(Chap.6),complexandadaptiveanalysisandsupporttasks(Chap.7) and self-referencing, self-awareness and self-interpretation within psychotherapy sessions (Chap. 8). Other chapters in Part III address mental models for learner- controlledlearning(Chap.9),attachmenttheory(Chap.12),controlledbondingby homophily(Chap.13)andsharedmentalmodelsinhospitalteams(Chap.14).Note that there is some overlap between the chapters concerning the general modelling approachused,inordertoallowforreadingthechaptersseparately. In addition, two chapters on philosophical perspectives on mental models are includedinPartIV.ThefirstofthemisChap.15,whichproposescontext-dependent realisationknownfromphilosophyofmindasaconceptualsolutionfortheissuethat different mental models may have different types of neural correlates. The second chapter from a philosophical perspective is Chap. 16, where it is shown how the approach to dynamics and adaptivity used here for mental models can be used to assign emerging informational content to mental models in a way that fits well in thegeneralapproachestodynamicsandadaptivitybasedontemporalfactorisation (Treur, 2007), relational specification of mental content (Kim, 1996), and criterial causation(Tse,2013). InPartV,somemoregeneralbackgroundisincludedonthemodellingapproach based on self-modelling networks used here: on the supporting software environ- ment(Chap.17),analysisofself-modellingnetworkbehaviourintermsofnetwork structure(Chap.18),validationofnetworkmodelsbyparametertuning(Chap.19) and mathematical analysis of the wide applicability of self-modelling networks to modelanyadaptivedynamicalsystem(Chap.20). InPartVIinChap.21,anevaluativeanalysisispresentedofwhattheformatof aself-modellingnetworkcanoffertothemodellingofmentalprocessesinvolving mentalmodels,coveringwhatalreadyhasbeenachievedandgeneratingoptionsthat still are open for further exploration. As an example, while the current book was alreadyinpress,foroneofthoseoptionsleftopeninthisbookithasbeenfoundout byGülayCanbalog˘luandJanTreurthattheapproachintroducedhereisveryuseful toobtaininasystematicmannercomputationalmodelsformultilevelorganisational learning,anareawhichlackedsuchcomputationalmodelsuntilnow.Anextbookby GülayandJan(togetherwithAnnaWiewiorafromtheareaofManagementScience) specificallyaddressingthisareabasedonthemodelingapproachintroducedinthe currentbookwillcomeoutlaterthisyearoratthebeginningofnextyear. Finally, we would like to thank Raj Bhalwankar, Gerrit Glas, Peter Roelofsma, Audrey Hermans, Selma Muhammad, and Jan Klein for their contribution to the developmentofthisinterestingresearchareaandasauthorsofoneormorechapters viii Preface tothisbookinparticular,andSefvanMentsforhishelpinthefinalproofreadingof thebook. Amsterdam,TheNetherlands JanTreur Jerusalem,Israel LailaVanMents November,2021 Contents PartI Introduction 1 Dynamics, Adaptation and Control for Mental Models: ACognitiveArchitecture ...................................... 3 LailavanMentsandJanTreur 2 BringingNetworkstotheNextLevel:Self-modelingNetworks forAdaptivityandControlofMentalModels .................... 27 JanTreur PartII Self-ModellingNetworkModelsforMentalModelsin IndividualProcesses 3 On Becoming a Good Driver: Modeling the Learning ofaMentalModel ............................................. 59 RajBhalwankarandJanTreur 4 Controlling Your Mental Models: Using Metacognition toControlUseandAdaptationforMultipleMentalModels ....... 81 JanTreur 5 DisturbedbyFlashbacks:AControlledAdaptiveNetwork ModelAddressingMentalModelsforFlashbacksfromPTSD ..... 99 LailavanMentsandJanTreur 6 ‘WhatifIWouldHaveDoneOtherwise…’:AControlled AdaptiveNetworkModelforMentalModelsinCounterfactual Thinking ..................................................... 117 RajBhalwankarandJanTreur 7 Do You Get Me: Controlled Adaptive Mental Models forAnalysisandSupportProcesses ............................. 139 JanTreur ix
The list of books you might like
