|
|
|
Chapitre d'ouvrage:
|
ID no. (ISBN etc.): 0-8058-3408-7
|
Clé de citation BibTeX: Holyoak2000
|
|
Holyoak, K. J., & Hummel, J. E. (2000). The proper treatment of symbols in a connectionist architecture. In E. Dietrich & A. B. Markman (Eds.), Cognitive dynamics: Conceptual and representational change in humans and machines. (pp. 229–263). Hillsdale, NJ: Erlbaum.
Ajoutée par: Sterenn Audo 2008-01-21 15:03:16 Dernièrement modifiée par: Sterenn Audo 2008-01-30 16:59:35
|
|
|
|
|
|
|
|
Résumé
from the book) The authors consider the relation between the physical symbol system hypothesis (A. Newell & H. A. Simon, 1976) and connectionist architectures with their so-called subsymbolic representations. Specifically, the authors consider the question "Can connectionist models eliminate physical symbol systems?" Their answer is "No." However, the authors find connectionist architectures to be powerful and important models, and their task is to show how connectionist models can incorporate symbols and symbol systems. They call this project symbolic connectionism. The authors begin by discussing in detail the necessity in cognitive modeling for variables and values (variable binding). They raise a problem for the tensor product view of variable binding. They then discuss the notion of a symbol distributed over space and time. Here they present their theory of synchrony-based variable bindings and their computer model of this theory: learning and inference with schemas and analogies (LISA). Finally, they show that symbolic connectionism is not merely implementation. (PsycINFO Database Record (c) 2007 APA, all rights reserved)
Ajoutée par: Sterenn Audo Dernièrement modifiée par: Sterenn Audo
|
|
Informations supplémentaires en ligne :
|
|
|
|
|
|
|
|
|
|
|
Idées
pdf dispo
Ajoutée par: Sterenn Audo
|
|
|
