Title :
Theory of symbol and pattern integration by associative memory with attention
Author_Institution :
Tokyo Univ. of Agric. & Technol., Japan
Abstract :
A theory based on the use of a metric is proposed as a background theory of a pattern and symbol integration method. A brain like structured associative memory model that has multiple metric fields and a switching mechanism is proposed as a neural implementation of the theory. A computer simulation shows a context sensitive association that is a base for natural language understanding
Keywords :
content-addressable storage; natural languages; neural nets; pattern recognition; sensor fusion; associative memory with attention; context sensitive association; metric fields; natural language understanding; neural implementation; switching mechanism; symbol/pattern integration; Agriculture; Associative memory; Brain modeling; Computer simulation; Humans; Inference algorithms; Motion control; Pattern recognition; Speech recognition; Sun;
Conference_Titel :
Multisensor Fusion and Integration for Intelligent Systems, 1996. IEEE/SICE/RSJ International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-3700-X
DOI :
10.1109/MFI.1996.572214
