Kernel recurrent system trained by real-time recurrent learning algorithm

Author

Pingping Zhu ; Principe, Jose C.

Author_Institution

Electr. & Comput. Eng., Univ. of Florida, Gainesville, FL, USA

fYear

2013

Firstpage

3572

Lastpage

3576

Abstract

This paper presents a kernelized version of recurrent systems (KRS) and develops a kernel real-time recurrent learning (KRTRL) algorithm to train KRS. To avoid instabilities during training, the teacher forcing technique is adopted to modify the KRTRL learning. The proposed algorithms compared with the KLMS in Lorenz time series prediction. The prediction performances of the proposed algorithm outperform the KLMS significantly.

Keywords

adaptive filters; learning systems; real-time systems; recurrent neural nets; time series; KRTRL algorithm; Lorenz time series prediction; kernel adaptive filter; kernel real-time recurrent learning algorithm; kernel recurrent system; teacher forcing technique; Algorithm design and analysis; Heuristic algorithms; Kernel; Prediction algorithms; Real-time systems; Signal processing algorithms; State-space methods; hidden state model; kernel adaptive filter; real-time recurrent learning (RTRL); recurrent networks; reproducing kernel Hilbert space (RKHS);

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on

Conference_Location

Vancouver, BC

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.2013.6638323

Filename

6638323

Link To Document

https://search.isc.ac/dl/search/defaultta.aspx?DTC=49&DC=1668184