Title :
Dynamic subgrouping in RTRL provides a faster O(N2) algorithm
Author :
Euliano, Neil R. ; Principe, Jose C.
Author_Institution :
Comput. Neuro Eng. Lab., Florida Univ., Gainesville, FL, USA
Abstract :
Static grouping of processing elements (PEs) has been proposed to reduce the computational complexity of real time recurrent learning (RTRL) from O(n4) to O(n2), but performance suffers. This paper proposes a dynamic subgrouping of PEs estimated from a local approximation of the π matrix based on temporal Hebbian of sensitivities during training. The method is O(n2) and leads to better performance
Keywords :
Hebbian learning; approximation theory; computational complexity; recurrent neural nets; π matrix; RTRL; computational complexity reduction; dynamic subgrouping; first-order approximation; local approximation; processing elements; real time recurrent learning; sensitivity matrix; temporal Hebbian learning; training; Computational complexity; Computer networks; Feedforward neural networks; History; Laboratories; Neural engineering; Neural networks; Recurrent neural networks; Sun; System identification;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2000. ICASSP '00. Proceedings. 2000 IEEE International Conference on
Conference_Location :
Istanbul
Print_ISBN :
0-7803-6293-4
DOI :
10.1109/ICASSP.2000.860135
