Title :
Address-data event representation for communication in multichip neuromorphic system architectures
Author :
Georgiou, J. ; Andreou, A.G.
Author_Institution :
Univ. of Cyprus, Nicosia
Abstract :
A new approach to asynchronous parallel-data conversion and transmission for computation in large neuromorphic distributed systems is reported. The proposed address data event representation (ADER) allows for efficient use of available bandwidth and ensures that no information is lost through collisions despite having a predominantly asynchronous nature.
Keywords :
microprocessor chips; address-data event representation; asynchronous parallel-data conversion; asynchronous parallel-data transmission; multichip neuromorphic system architectures;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:20070897
