Abstract
States or state sequences in neural network models are made to represent concepts from applications. This paper motivates, introduces and discusses a formalism for denoting such representations; a representation for representations. The formalism is illustrated by using it to discuss the representation of variable binding and inference abstractly, and then to present four specific representations. One of these is an apparently novel hybrid of phasic and tensor-product representations which retains the desirable properties of each.
Original language | English |
---|---|
Title of host publication | Neurodynamics and Psychology |
Editors | M. Oaksford, G. D. A. Brown |
Publisher | Academic Press |
Pages | 255-266 |
Number of pages | 12 |
ISBN (Print) | 0125235151 |
Publication status | Published - 4 Jan 1994 |
Event | Neuro Dynamics and Psychology - Duration: 4 Jan 1994 → 4 Jan 1994 |
Other
Other | Neuro Dynamics and Psychology |
---|---|
Period | 4/01/94 → 4/01/94 |
Bibliographical note
Error in formula 25.Keywords
- neural network
- representation
- variable binding
- inference abstractly