Title :
Distributed probability network
Author_Institution :
Dept. of Electr. & Comput. Eng., San Diego State Univ., CA, USA
fDate :
6/22/1905 12:00:00 AM
Abstract :
This paper introduces a distributed sub-pattern space binary associative memory for error correction and pattern completion. The associative architecture presented in this work is analogous to belief network representations in that, it represents domain knowledge using a set of sub-pattern joint probability distributions. It is superior to belief networks in that, it can be used as an associative memory. The proposed architecture is distributed in nature and offers a fault tolerant solution to binary pattern completion and error correction problems
Keywords :
belief networks; content-addressable storage; distributed memory systems; error correction; neural nets; probability; artificial neural networks; associative architecture; belief network representations; binary associative memory; binary error correction; binary pattern completion; biological memory systems; distributed architecture; distributed probability network; distributed sub-pattern space memory; domain knowledge; fault tolerant solution; learning; neural network; sub-pattern joint probability distributions; Artificial neural networks; Associative memory; Biology computing; Computer architecture; Degradation; Error correction; Fault tolerance; Information retrieval; Memory architecture; Probability distribution;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on
Print_ISBN :
0-7803-6293-4
DOI :
10.1109/ICASSP.2000.860130
