Are feedforward and recurrent networks systematic? Analysis and implications for a connectionist cognitive architecture
Philosophical Explorations (1998)
Human cognition is said to be systematic: cognitive ability generalizes to structurally related behaviours. The connectionist approach to cognitive theorizing has been strongly criticized for its failure to explain systematicity. Demonstrations of generalization notwithstanding, I show that two widely used networks (feedforward and recurrent) do not support systematicity under the condition of local input/output representations. For a connectionist explanation of systematicity, these results leave two choices, either: (1) develop models capable of systematicity under local input/output representations; or (2) justify the choice of similarity-based (nonlocal) component representations sufficient for systematicity.
|Keywords||No keywords specified (fix it)|
|Categories||categorize this paper)|
References found in this work BETA
No references found.
Citations of this work BETA
Connectionist Semantic Systematicity.Stefan L. Frank, Willem F. G. Haselager & Iris van Rooij - 2009 - Cognition 110 (3):358-379.
Similar books and articles
Strong Semantic Systematicity From Hebbian Connectionist Learning.Robert F. Hadley & M. B. Hayward - 1997 - Minds and Machines 7 (1):1-55.
Systematic Minds, Unsystematic Models: Learning Transfer in Humans and Networks. [REVIEW]Steven Phillips - 1999 - Minds and Machines 9 (3):383-398.
Connectionism and Cognitive Architecture.Jerry A. Fodor & Zenon W. Pylyshyn - 1988 - Cognition 28 (1-2):3-71.
The Allure of Connectionism Reexamined.Brian P. McLaughlin & F. Warfield - 1994 - Synthese 101 (3):365-400.
To Transform the Phenomena: Feyerabend, Proliferation, and Recurrent Neural Networks.Paul M. Churchland - 1997 - Philosophy of Science 64 (4):420.
Added to index2009-01-28
Total downloads11 ( #400,729 of 2,164,288 )
Recent downloads (6 months)1 ( #348,039 of 2,164,288 )
How can I increase my downloads?