Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of Alberta

$24,745.00

Artificial neural networks, contingency, and associative learning

EF 2009-10

Résumé
Cognitive science adopts the tri-level hypothesis, which directs researchers to address three different questions about cognition: what information processing problem is being solved? What information processing steps are required to solve the problem? What specific mechanisms are required to execute these information processing steps? The tri-level hypothesis has also been applied to associative learning, which attempts to explain how we remember relationships between ideas and events. Shanks has argued that the implementational question about associative learning is answered by artificial neural networks, which are computer simulations of brain-like information processing. The proposed research is partly motivated by a concern with Shanks' position: my previous research has shown that artificial neural networks do not merely provide implementational answers. Rather, artificial neural networks provide insights at both the computational and at the algorithmic levels. Recently, I have used artificial neural networks to study a simple form of associative learning called classical or Pavlovian conditioning. This work - which involved both formal proofs and computer simulations of learning experiments - revealed a number of fundamental issues about the relationship between artificial neural networks and associative learning. My future research will extend this research paradigm to deal with more complex forms of learning - the learning of the contingencies between events, and to associative learning in general. I expect that my work on artificial neural networks will lead to new theories of associative learning, and will clarify the theoretical and empirical relationships between networks and theories of association.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Alberta, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
2038-2007
Name nom
Dawson, Michael
Department département
Psychology
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2009
Competition year année de concours
2007
Award amount
24745
Program id
RGPIN
Program nane en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
12
Committee name en
Psychology: Brain, Behaviour & Cognitive Sciences
Committee name fr
Psychologie: Cerveau, comportement et science cognitive
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
6304
Research subject group en
Psychology
Research subject group fr
Psychologie
Research subject en
Cognitive science -- fundamental
Research subject fr
Science cognitive -- principes fondamentaux
Partie
3
Nb partie
5
Application title
Artificial neural networks, contingency, and associative learning
Keyword
NA
Application summary
Cognitive science adopts the tri-level hypothesis, which directs researchers to address three different questions about cognition: what information processing problem is being solved? What information processing steps are required to solve the problem? What specific mechanisms are required to execute these information processing steps? The tri-level hypothesis has also been applied to associative learning, which attempts to explain how we remember relationships between ideas and events. Shanks has argued that the implementational question about associative learning is answered by artificial neural networks, which are computer simulations of brain-like information processing. The proposed research is partly motivated by a concern with Shanks' position: my previous research has shown that artificial neural networks do not merely provide implementational answers. Rather, artificial neural networks provide insights at both the computational and at the algorithmic levels. Recently, I have used artificial neural networks to study a simple form of associative learning called classical or Pavlovian conditioning. This work - which involved both formal proofs and computer simulations of learning experiments - revealed a number of fundamental issues about the relationship between artificial neural networks and associative learning. My future research will extend this research paradigm to deal with more complex forms of learning - the learning of the contingencies between events, and to associative learning in general. I expect that my work on artificial neural networks will lead to new theories of associative learning, and will clarify the theoretical and empirical relationships between networks and theories of association.