TY - GEN
T1 - An analog network approach to train RBF networks based on sparse recovery
AU - Feng, Ruibin
AU - Leung, Chi-Sing
AU - Constantinides, A. G.
PY - 2014
Y1 - 2014
N2 - The local competition algorithm (LCA) is an analog neural approach for compressed sensing. It is used to recover a sparse signal from a set of measurements. Unlike some traditional numerical methods that produce many elements with small magnitude, the LCA automatically set those unimportant elements to zero. This paper formulates the training process of radial basis function (RBF) networks as a compressed sensing problem. We then apply the LCA to train RBF networks. The proposed LCA-RBF approach can select important RBF nodes during training. Since the proposed approach can limit the magnitude of the trained weight, it also has certain ability to handle RBF networks with multiplicative weight noise.
AB - The local competition algorithm (LCA) is an analog neural approach for compressed sensing. It is used to recover a sparse signal from a set of measurements. Unlike some traditional numerical methods that produce many elements with small magnitude, the LCA automatically set those unimportant elements to zero. This paper formulates the training process of radial basis function (RBF) networks as a compressed sensing problem. We then apply the LCA to train RBF networks. The proposed LCA-RBF approach can select important RBF nodes during training. Since the proposed approach can limit the magnitude of the trained weight, it also has certain ability to handle RBF networks with multiplicative weight noise.
KW - Fault tolerance
KW - Local competition algorithm
KW - RBF networks
UR - https://www.scopus.com/pages/publications/84940758462
UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-84940758462&origin=recordpage
U2 - 10.1109/ICDSP.2014.6900799
DO - 10.1109/ICDSP.2014.6900799
M3 - RGC 32 - Refereed conference paper (with host publication)
SN - 9781479946129
VL - 2014-January
SP - 903
EP - 908
BT - International Conference on Digital Signal Processing, DSP
PB - IEEE
T2 - 19th International Conference on Digital Signal Processing (DSP 2014)
Y2 - 20 August 2014 through 23 August 2014
ER -