Monthly
288 pp. per issue
6 x 9, illustrated
ISSN
0899-7667
E-ISSN
1530-888X
2014 Impact factor:
2.21

Neural Computation

December 2012, Vol. 24, No. 12, Pages 3246-3276
(doi: 10.1162/NECO_a_00374)
© 2012 Massachusetts Institute of Technology
Design Strategies for Weight Matrices of Echo State Networks
Article PDF (708.03 KB)
Abstract

This article develops approaches to generate dynamical reservoirs of echo state networks with desired properties reducing the amount of randomness. It is possible to create weight matrices with a predefined singular value spectrum. The procedure guarantees stability (echo state property). We prove the minimization of the impact of noise on the training process. The resulting reservoir types are strongly related to reservoirs already known in the literature. Our experiments show that well-chosen input weights can improve performance.