Quantum Cognition: A Cognitive Architecture For Human-AI Thiab In-Memory Computing Part 1
Nov 07, 2023
Kab lus no tsom mus rau human-artificial txawj ntse (AI): "cov tshuab uas xav, uas kawm, thiab tsim". Kuv pom qhov teeb meem ntawm qee qhov teeb meem uas tau ua rau tsis muaj kev vam meej hauv AI (kev vam meej ntau dua hauv kev txawj ntse thiab tsis muaj kev txawj ntse) thiab qhia quantum cognition asa siv tau kev txawj ntse architecture rau tib neeg AI thiab kho vajtse.
Nyob rau hauv xyoo tas los no, nrog rau kev txhim kho sai ntawm kev txawj ntse technology, ntau thiab ntau tus neeg tau pib xav txog kev sib raug zoo ntawm kev txawj ntse thiab kev nco. Kev sib raug zoo ntawm ob leeg yog inseparable.
Artificial txawj ntse muaj peev xwm simulate tib neeg lub cim xeeb txheej txheem, cia cov cav tov nco ntsoov cov ntaub ntawv thiab khaws cia thiab retrieve nws zoo li tib neeg. Hauv qhov kev nkag siab no, thev naus laus zis kev txawj ntse tau nthuav dav tib neeg lub peev xwm nco. Piv txwv li, kev txawj ntse txawj ntse tuaj yeem pab peb nrog kev hais lus thiab paub cov duab, yog li pab peb nco thiab nkag siab cov ntaub ntawv zoo dua.
Ntawm qhov tod tes, tib neeg lub peev xwm nco tau tuaj yeem muab kev tshoov siab tseem ceeb rau kev txhim kho kev txawj ntse. Tib neeg lub peev xwm nco tau tuaj yeem pab peb tswj hwm thiab txheeb xyuas cov ntaub ntawv zoo dua thiab nrhiav kev koom tes thiab cov qauv ntawm cov ntaub ntawv. Cov peev txheej no tseem tuaj yeem hloov mus rau hauv algorithms thiab cov kev pab cuam los pab cov khoos phis tawj ua haujlwm zoo dua thiab siv cov ntaub ntawv. Yog li ntawd, tib neeg lub peev xwm nco tau yog qhov tseem ceeb rau kev txhim kho kev txawj ntse.
Hauv luv luv, kev sib raug zoo ntawm kev txawj ntse txawj ntse thiab kev nco yog kev sib koom ua ke. Lawv txhawb ib leeg thiab sib koom ua ke txhawb kev siv technology thiab kev vam meej. Cia peb ntsib txoj kev loj hlob ntawm kev txawj ntse txawj ntse thiab kev nco zoo, nrog tus cwj pwm zoo, thiab ntseeg tias lawv tuaj yeem coj tau lub neej yav tom ntej zoo dua. Nws tuaj yeem pom tau tias peb yuav tsum txhim kho peb lub cim xeeb, thiab Cistanche deserticola tuaj yeem pab peb txhim kho peb lub cim xeeb vim Cistanche deserticola yog cov khoom siv tshuaj suav tshuaj uas muaj ntau yam tshwj xeeb, ib qho ntawm kev txhim kho kev nco. Qhov ua tau zoo ntawm cov nqaij minced yog los ntawm ntau cov khoom xyaw uas nws muaj, nrog rau cov kua qaub, polysaccharides, flavonoids, thiab lwm yam. Cov khoom xyaw no tuaj yeem txhawb lub hlwb kev noj qab haus huv ntau yam.
Nyem paub 10 txoj hauv kev los txhim kho kev nco
Feem ntau ntawm cov kev txawj ntse architectures, uas yog, cov qauv ntawm tib neeg kev xav hauv kev tshawb fawb txog kev txawj ntse (AI), tsis tas yuav sim tomodel cov txheej txheem tib neeg kev xav.Lawv xav tias tib neeg yog cov neeg muaj peev xwm, uas yog, cov khoom siv hluav taws xob, uas ib txwm ua raws li Boolean logic, uas yog qhov tseeb tias. Cov xwm txheej ib txwm tuaj yeem sib xyaw ua ke (piv txwv li, ntawm kev sib txuas lus) hauv txhua qhov kev txiav txim. Lawv sim ntsuas qhov sib npaug ntawm chav ntsuas arithmeticlogic (ALU) hauv tib neeg lub hlwb, thiab sib txuas cov ntaub ntawv los ua kom nws ywj pheej thiab sib faib (IID). Cov hauv qab no qhia cov ntsiab lus ntxiv.
TSEEM CEEB nrog CLASSICALPROBABILITY INREASONING
Classical probability theory and widely accept Kolmogorov axiomsfollow Boolean logic. Qhov no txhais tau hais tias lub logic ntawm cov xwm txheej yog kev sib hloov thiab cov xwm txheej yeej ib txwm sib xws. Ntawd yog, A thiab B yog tib yam B thiab A, thiab kev ntsuas ib txhij ntawm A thiab B, lossis B thiab A yuav tsis cuam tshuam.
Cov logic no ua haujlwm zoo rau cov xwm txheej sib xws. Piv txwv li, ua ntej ntsuas koj qhov siab thiab tom qab ntawd koj qhov hnyav, ua ntej ntsuas koj qhov hnyav thiab qhov siab, lossis ib txhij ntsuas koj qhov siab thiab qhov hnyav tag nrho cov txiaj ntsig zoo ib yam.
Tab sis qhov tseeb yog tias cov xwm txheej yuav tsis sib haum, uas yog, kev ntsuam xyuas yog nyob ntawm kev txiav txim, thiab cuam tshuam tuaj yeem tshwm sim. Piv txwv li, xav txog questionA: "Koj puas mus Florida?" thiab nug B: "Koj puas hnov muaj cua daj cua dub los rau Florida?" Ua ntej thov cov lus teb rau lo lus nug A thiab tom qab ntawd nug B, lossis thawj zaug thov cov lus teb rau lo lus nug B thiab tom qab ntawd nug A, lossis nug cov lus nug A thiab B ib txhij thiab tom qab ntawd thov cov lus teb yuav ua rau cov lus teb sib txawv.
Tsis tas li ntawd, ib qho kev sib txuas ntawm Boolean tuaj yeem pom tau tias yog tus neeg sawv cev ntau dua li ib qho ntawm nws cov neeg sawv cev thiab hloov tib neeg kev xav. Ntawm no yog ib qho piv txwv yooj yim.
Xav tias Bob tau raug txheeb xyuas tias yog tus neeg raug liam ntawm kev siv lub sijhawm xoom-hnub. Tsis tas li ntawd, xav tias kev siv qhov tsis zoo no feem ntau tau pom los ntawm cov tswv cuab ntawm pawg neeg nyiag nkas npe hu ua H.Tom qab ntawd, qhov twg ntawm cov xwm txheej hauv qab no yuav tshwm sim ntau dua?
1. Bob yog tus kws tshaj lij hacker.
2. Bob yog tus kws tshaj lij hacker thiab amember ntawm pawg H.
Intuitively hais lus, scenario 2 tuaj yeem pom tau tias yog qhov tshwm sim ntau dua.Tab sis nrog Boolean logic thiab classicalprobability txoj kev xav, qhov tshwm sim ntawm ob qhov xwm txheej tshwm sim ua ke tsis tuaj yeem loj dua qhov tshwm sim ntawm asingle tshwm sim. Peb pom tias qhov xwm txheej 2 yog qhov tshwm sim ntau dua li qhov xwm txheej 1 vim yog kev sib txuas lus tsis zoo, kev paub tsis meej uas tau txheeb xyuas los ntawm Tversky thiab Kahneman1 uas piav qhia tias tib neeg feem ntau xav ntseeg ntau zaj dab neeg nrog cov ntsiab lus nthuav dav tshaj qhov luv luv. Phishing attackers tau txais txiaj ntsig zoo los ntawm qhov kev tsis ncaj ncees no los ntawm thawj zaug muab lawv lub hom phiaj nrog cov lus piav qhia meej thiab nthuav dav ntawm qhov xwm txheej uas yuav tsum tau saib xyuas tam sim ntawd thiab nug lawv kom nyem qhov txuas.
KEV NTSEEG KEV KAWM
Goyal thiab Bengio2 sib cav hais tias kom ua tiav tib neeg AI, peb yuav tsum tau txav mus los ntawm qhov system 1 / implicit / tsis nco qab ua tiav mus rau qhov system 2 / meej / meej meej ua. System 1 kev ua haujlwm zoo ib yam li thaum peb tsav tsheb hauv cov neeg nyob ze, uas peb tuaj yeem ceev ceev thiab tsis nco qab. Kev ua haujlwm 2 zoo ib yam li thaum peb tsav tsheb hauv ib cheeb tsam uas tsis paub thiab xav tau qeeb thiab nco qab thiab xav tau kev sab laj ib yam.
Goyal thiabBengio lub tswv yim yuav tsum "ua raws li kev nco qab" thiab xav txog "kev saib xyuas raws li kev xaiv kev suav ua dab tsi los ua qhov ntau." Txawm li cas los xij, raws li tau tham luv luv, qhov yuav tshwm sim classical muaj qhov txwv loj nrog kev ua ntu zus.Nws xav tias txhua yam xwm txheej tau sib xws thiab tsis xav txog kev txiav txim.
Piv txwv li, kom tsis txhob overfitting (hloov ntau dhau rau cov ntaub ntawv tshwj xeeb nws tau kawm), lub tshuab kev kawm hauv zej zog shufflesdata ua rau lawv IID. Tab sis qhov tseeb yog cov ntaub ntawv tsis tuaj txog peb li IID.2
"Nature tsis shuffle cov ntaub ntawv, thiab peb yuav tsum tsis txhob. Thaum weshuffle cov ntaub ntawv, peb rhuav tshem cov ntaub ntawv hais txog cov kev hloov nyob rau hauv tis uas muaj nyob rau hauv cov ntaub ntawv wecollect thiab muaj cov ntaub ntawv hais txog cov qauv causal."
QUANTUM PROBABILITY RAU REASONINGAND INFERENCE
Kuv pom zoo kom quantum cognition3 yog ib txoj hauv kev siv tau rau kev txawj ntse architectures uas siv classical vim li cas thiab inference. Quantum cognition txawv ntawm quantummind. Nws tsis ua raws li qhov kev xav tias muaj ib yam dab tsi quantum zoo li tshwm sim hauv lub hlwb tab sis kev tshoov siab los ntawm kev ua lej ntawm quantum txoj kev xav thiab nws cov ntsiab lus dynamic. Piv txwv li, nws siv quantum probability-modelingcognition siv txoj kev xav ntawm qhov tshwm sim los ntawm quantum mechanics, yam tsis muaj physics.
Cov lus hauv qab no qhia txog qhov piv txwv ntawm quantum probability uas ua rau nws tsim nyog rau tib neeg-AI software thiab kho vajtse.
CAPTURINGINCOMPATIBILITIES
Quantum probability, tsis zoo li classicalprobability, xav tias tag nrho cov lus nug tau sib xws, thiab tuaj yeem ntes qhov tsis sib xws. Quantum probability sivvector qhov chaw thiab subspace zoo ib yam li classical probability qhov kev siv ntawm samplespace thiab kev tshwm sim (uas yog, ib tug subset ntawm piv txwv qhov chaw), feem. Vectorspace muaj tag nrho cov txiaj ntsig tau rau cov lus nug. Ib lub vector uas sawv cev rau cov lus nug tshwm sim nthuav tawm 1D subspace, hu ua kab hluav taws xob, thiab cov txheej txheem kev ntseeg uas tus neeg muaj txog cov lus nug uas sawv cev los ntawm chav tsev-ntev vector, hu ua lub xeev vector. Quantum probability kuj tseem siv cov txheej txheem kev kos duab, hu ua projecting, thiab qhov tshwm sim tau muab rau ib qho xwm txheej sib npaug ntawm qhov ntev ntawm qhov projection. Tocompute qhov kev sib txuas ntawm cov lus nug tshwm sim, quantum probability ntiav ib qho kev sib txuas ua ke. Qhov no tso cai rau qhov sib txawv ntawm qhov kev txiav txim, uas yog, qhov project A thiab tom qab ntawd qhov project Bhas qhov sib txawv ntawm qhov projectB thiab tom qab ntawd qhov project A.
Rov xyuas qhov piv txwv yooj yim
Ntawm no, peb rov mus saib peb qhov piv txwv yooj yim los piav qhia seb yuav ua li cas quantum probability, siv vector space, tuaj yeem piav qhia txog kev sib koom ua ke hauv tib neeg kev xav. Nyob rau hauv daim duab 1, xiav xiav qhia "Bob ua tus kws tshaj lij hacker" los ntawm B thiab nws qhov kev tsis pom zoo nrog ~ / B. Ib yam li ntawd, cov xib xub txiv kab ntxwv sawv cev rau "ua tus tswv cuab ntawm pawg H" nrog H thiab nws tsis pom zoo nrog ~ / H. S, lub xeev vector, sawv cev rau peb txoj kev ntseeg ntawm Bob's characterization thiab sawv cev los ntawm cov xub dub. Nyob rau hauv daim duab 1, txoj kev projection yog qhia los ntawm ntsuab andred dotted kab.
Qhov yuav tshwm sim yog xam raws li qhov ntev squared ntawm qhov projection ntawm lub xeev vector mus rau qhov sib thooj axis thiab qhia los ntawm ntsuab thiab liab square lengths. Qhov projection mus rau B ray yog qhia los ntawm cov kab ntsuab dotted, thiab qhov tshwm sim ntawm (B) sib npaug ntawm qhov squared length ntawm qhov bar, qhia los ntawm ntsuab squarelength. Rau qhov tshwm sim ntawm (B thiab H), peb yuav tsum ua raws li ob kauj ruam, pom los ntawm ob kab liab dotted. Ua ntej, peb npaj lub xeev vector mus rau H ray. Qhov thib ob, peb ua qhov project yav dhau los no mus rau B ray. Tom qab ntawd, qhov tshwm sim ntawm (B thiab H) yog qhov ntev squared ntawm qhov kawg qhov projection, qhia los ntawm qhov ntev square liab.
Nyob rau hauv daim duab 1, qhov tshwm sim tshwm sim ntawm (B thiab H) yog ntau dua qhov tshwm sim ntawm ib qho kev tshwm sim, uas yog, qhov tshwm sim ntawm (B), sib raug rau qhov ntev square liab ntev dua qhov ntev ntawm ntsuab square. Qhov no yog vim qhov kev sib txuas tsis sib xws uas ua rau pom qhov xwm txheej 2 yog qhov tshwm sim ntau dua li qhov xwm txheej 1. Wecan cuam tshuam txog qhov tsis sib xws ntawm (B thiab H), ua rau lawv cuam tshuam, rau qhov muaj peev xwm heuristic (ib qho kev qhia luv luv); Kev sib txuas zoo li tus neeg sawv cev ntau dua li ib qho ntawm nws cov neeg sawv cev, thiab ua ib tus tswv cuab ntawm H tuaj yeem yooj yim rau kev xav lossis kom rov qab tau dua li Bob ua ib pawg. Rau kev piav qhia lej ntawm qhov piv txwv no, saib cov ntaub ntawv ntxiv muaj nyob ntawm 10.1109/MC.2023.3242056.
Lub peev xwm ntawm quantum probability to capture incompatibilities kuj tseem ua lub luag haujlwm tseem ceeb hauv kev tsim cov qauv tsim rau tib neeg AII, tshwj xeeb tshaj yog thaum peb cuam tshuam nrog cov xwm txheej tsis sib xws los ntawm kev sib koom ua ke nrog cov xwm txheej loj heev ntawm cov ntaub ntawv los ntawm ntau qhov chaw. Nyob rau hauv cov xwm txheej zoo li no, peb xav tau cov qauv tsim qauv tsim qauv los nthuav tawm cov txheej txheem hauv qab ntawm cov ntaub ntawv piv rau lub hauv paus causal induction, uas yog, modelinga ib qho kev sib raug zoo-thiab-zoo, siv classical probability. Nyob rau hauv cov xwm txheej zoo li no, quantum probability tuaj yeem muab txoj hauv kev los tsim lub tswv yim ntawm cov qauv hauv zos qhov laj thawj los ntawm kev ua haujlwm nrog cov xwm txheej tsis sib xws, pasting ua ke cov qauv qhov chaw, thiab tsim ib qhov chaw vector.
Piv txwv li, xav tias peb xav tau kev txiav txim siab kwv yees, uas yog, nrhiav qhov xwm txheej tshwm sim ntawm qhov tshwm sim vim li cas, lossis P (effect|ua rau), hauv cov teeb meem nyuaj nrog cov ntaub ntawv loj heev, qhov kev txiav txim ntawm cov ntaub ntawv tuaj txog. Quantum probability enablesus kom tawg qhov teeb meem mus rau hauv me me teeb meem los ntawm teb cov lus nug xws li:
P (effect|ua 1, tsis muaj lwm txoj hauv kev), P (effect|ua rau 1, ua rau 2), P (effect|ua rau 2, ua rau 1), thiab lwm yam.
QUANTUM PROBILITY FORIN-MEMORY COMUTING
Quantum probability siv vectorspace, zoo ib yam li kev siv lub moj khaum vector symbolic architectures (VSAs), tseem hu ua hyperdimensional xam, uas yog lub hauv paus rau cov khoom siv tshiab, piv txwv li, hauv-nco xam (IMC). Nyob rau hauv cov pa von Neumann architecture, lub cim xeeb thiab processor yog cais, thiab cov kev suav yuav tsum tau tsiv mus nyob rau hauv thiab tawm. Tab sis nrog IMC architecture siv "vector-matrix multiplication,"4 nco thiab processor yog fused, thiab kev xam tau ua qhov twg cov ntaub ntawv khaws cia nrog cov ntaub ntawv tsawg tsawg. Uas ua rau IMC, nyob rau hauv sib piv rau cov pa von Neumann architecture, zoo ib yam li tib neeg lub hlwb, wherememory thiab kev suav yog collocated. Nrhiav qhov sib npaug ntawm ALU hauv tib neeg lub hlwb yog qhov kev cia siab tsis muaj tseeb.
Quantum probability thiab IMC tuaj yeem suav hais tias yog kev cog lus suav nrog cov qauv tsim rau tib neeg AII raws li siv VSAs. Yog li nws yog qhov tsim nyog los xav txog quantum probability raws li kev txawj ntse architecture rau IMC.
Ntawm no yog ib qho piv txwv. Ua hauj lwm nco nyob rau hauv tib neeg lub hlwb yog ib tug txheej txheem rau ib ntus cia ntawm cov ntaub ntawv ntsig txog cov hauj lwm tam sim no. Nws yog ib qho tseem ceeb rau kev txawj ntse xws li kev mloog, kev xav, thiab kev kawm; Yog li, feem ntau kev txawj ntse architectures siv nws nyob rau hauv ib co daim ntawv.With quantum cognition, peb muaj peev xwm siv high-dimensional vectors los sawv cev lub luag hauj lwm ntawm kev ua hauj lwm nco thiab todeal nrog rau cov ntaub ntawv nyob rau hauv ib tug tsis tu ncua kev xam. Quantum probability lub xeev vector tuaj yeem suav tias yog lub xeev kev nco ua haujlwm uas sawv cev rau tib neeg kev ntseeg txog cov qauv qauv thiab ua haujlwm raws li lub cache rau cov qauv hauv ntiaj teb tam sim no, lub xeev ntawm lub cev, thiab / lossis cov hom phiaj tam sim no. kev ua haujlwm ntawm cov qauv hyperdimensional uas tuaj yeem siv rau kev suav." Los ntawm kev saib cov qauv raws li vectors, peb tuaj yeem "kais mus rau hauv lub cev loj ntawm kev paub txog vectors, matrices, linear algebra, thiab dhau mus. Qhov no muaj tseeb yog qhov kev lig kev cai hauv kev tshawb fawb neural-net, tab sis cov chaw nplua nuj ntawm cov sawv cev siab tseem yuav tau tshawb nrhiav." Cov neeg sawv cev uas ib txwm ua kom muaj txiaj ntsig zoo rau nws, peb yuav tsum ua raws li Booleanlogic, yog qhov tseem ceeb rau AIcomputing architectures uas twb muaj lawm. Tab sis kuv sib cav tias peb yuav tsum kawm los ntawm Einstein txoj kev ntseeg sib txheeb, Gödel's incompleteness theorem, thiab Simon's bounded rationality theory, raws li lawv txhua tus ua rau pom qhov tawg ntawm qhov tseeb.
Hauv tsab xov xwm no, kuv tau nthuav tawm qee qhov kev txwv kev suav ntawm AIsystems uas twb muaj lawm. Kuv piav qhia tias, tsis zoo li cov axioms ntawm classical probability, lub logic ofevents tsis tas Boolean. Yog tias ob qhov xwm txheej A thiab B tsis sib haum, ces kawm; Yog li, feem ntau kev txawj ntse architectures siv nws nyob rau hauv ib co daim ntawv.With quantum cognition, peb muaj peev xwm siv high-dimensional vectors los sawv cev lub luag hauj lwm ntawm kev ua hauj lwm nco thiab todeal nrog rau cov ntaub ntawv nyob rau hauv ib tug tsis tu ncua kev xam. Quantum probability lub xeev vector tuaj yeem suav hais tias yog lub xeev kev nco ua haujlwm uas sawv cev rau cov kev ntseeg ntawm cov qauv thiab ua haujlwm raws li lub cache rau cov qauv hauv ntiaj teb tam sim no, lub xeev ntawm lub cev, thiab / lossis cov hom phiaj tam sim no.
Quantum probability tsim lub hauv paus rau kev ua lej muaj zog rau IMC thiab teeb tsa "kev ua haujlwm ntawm cov qauv hyperdimensional uas tuaj yeem siv rau kev suav." Los ntawm kev saib cov qauv ua vectors, peb tuaj yeem ua tau
"kais mus rau hauv lub cev loj ntawm kev paub txog vectors, matrices, linear algebra, thiab dhau mus. Qhov no muaj tseeb yog qhov kev lig kev cai hauv kev tshawb fawb neural-net, tab sis cov chaw nplua nuj ntawm cov sawv cev siab tseem yuav tau tshawb nrhiav."
Cov neeg sawv cev uas ib txwm ua kom muaj txiaj ntsig zoo rau nws, peb yuav tsum ua raws li Booleanlogic, yog qhov tseem ceeb rau AI uas twb muaj lawm qhov kev sib txuas ntawm cov xwm txheej A thiab B tsis tuaj yeem txhais tau vim tias lawv tsis mus, hauv qhov sib piv nrog Booleanlogic, qhov xwm txheej ib txwm mus. .
Kuv tau muab cov lus pom zoo txog qhov tshwm sim ntawm quantum thiab piav qhia yuav ua li cas los txiav txim siab quantum lub xeev kev ntsuas ntawm cov qauv uas tsis yog Boolean ntawm cov neeg ua haujlwm projection. Txhawm rau sib piv cov xeev quantum thiab cov xeev classicalprobabilistic, kuv tau piav qhia yuav ua li cas qhov projection tuaj yeem siv los piav qhia cov kev sim zoo ib yam li classical probability. Kuv tau piav qhia yuav ua li cas cov qauv tsim qauv (piv rau lub hauv paus causalinduction) tuaj yeem pab nrog kev ua tiav kev nco qab. Ialso tau piav qhia yuav ua li cas quantum probability qhov kev siv ntawm vector qhov chaw ua rau nws tsis tsim nyog kev txawj ntse architecture rauIMC architecture.
Ua tiav tib neeg AII thiab tsim "tshuab uas xav tias kawm thiab tsim" 6 xav tau cov qauv kev suav uas tuaj yeem ua tau zoo ib yam. Tab sis tib neeg txoj kev xav, kev kawm, thiab kev tsim feem ntau yog cov ntsiab lus tseem ceeb thiab kev txiav txim siab, thiab qhov ntawd zoo li tsis txaus ntseeg rau qhov ua tau zoo tshaj plaws thiab cov qauv siv ntau tshaj plaws.
TSEEM CEEB
Cov ntaub ntawv no yog raws li kev ua haujlwm txhawb nqa los ntawm National ScienceFoundation nyob rau hauv Award 2041788. Cov lus piav qhia luv luv ntawm kev ua lej ntawm qhov kev xeem yooj yim thiab nws rov tuaj xyuas tuaj yeem nrhiav tau hauv cov ntaub ntawv ntxiv muaj nyob ntawm 10.1109/MC.2023.3242056.
REFERENCES
1. A. Tversky thiab D. Kahneman, "Extensional versus intuitive reasoning: The conjunction fallacy inprobability judgment," Psychol. Rev., vol. 9 0, nr. 4, pp. 293–315, Nov. 1983, doi: 10.1037/0033-295X.90.4.293.
2. A. Goyal thiab Y. Bengio, "Inductive biases rau kev kawm tob ntawm kev paub ntau dua," Proc. Roy. Soc. A, vol. 478, Nr. 2266, Nov. 2022, Art. tsis muaj. 20210068, doi: 10.1098/rpa.2021.0068.
3. JR Busemeyer thiab P. Bruza, Quantum Models of Cognition and Decision.Cambridge, UK: Cambridge Univ.Press, 2014.
4. S. Spetalnick thiab A. Raychowdhury, "Ib tug tswv yim tsim-chaw tsom xam ntawm xam-hauv-nco nrogSRAM," IEEE Trans. Lub Circuit Court Syst. I, Reg. Cov ntaub ntawv, vol. 6 9, nr. 4, pp. 1466–1479, Apr. 2022, doi: 10.1109/TCSI.2021.3138057.
5. P. Kanerva, "Hyperdimensionalcomputing: Ib qho kev taw qhia tocomputing nyob rau hauv faib sawv cev nrog high-dimensional randomvectors," Cogn. Comput., ib. 1, p. 139–159, Peb. 2009, doi: 10.1007/s12559-009-9009-8.
6. H. Simon thiab A. Newell, "Heuristicproblem solving: The next advance inoperations research," Oper. Res., vol. 6, nre. 1, p. 1–10, Peb./Feb. Xyoo 1958. [Online].Available:https://www.jstor.org/stable/167397
For more information:1950477648nn@gmail.com
