Feedback Error Learning with a noisy teacher

Author

Ishihara, Abraham K. ; van Doornik, Johan ; Ben-Menahem, Shahar

Author_Institution

Carnegie-Mellon Univ., Moffett Field, CA

fYear

2008

fDate

11-13 June 2008

Firstpage

4529

Lastpage

4534

Abstract

The feedback error learning (FEL) algorithm is examined under the condition of a noisy teaching signal. The teaching signal, which adaptively adjusts the weights of the feedforward network, is assumed to be corrupted by a signal dependent noise (SDN) source. The FEL framework was originally inspired by the cerebellum as a model for human motor control. We analyze the robustness properties of the original system with respect to the SDN noise model. We prove bounds on the learning rate and feedback gain matrices that guarantee stochastic stability of the closed loop system.

Keywords

adaptive control; closed loop systems; feedback; neurocontrollers; robust control; stochastic systems; closed loop system; feedback error learning algorithm; feedback gain matrices; feedforward network; human motor control; noisy teacher; noisy teaching signal; signal dependent noise source; stochastic stability; Brain modeling; Control system analysis; Education; Feedback loop; Humans; Motor drives; Noise robustness; Stability; Stochastic resonance; Stochastic systems;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 2008

Conference_Location

Seattle, WA

ISSN

0743-1619

Print_ISBN

978-1-4244-2078-0

Electronic_ISBN

0743-1619

Type

conf

DOI

10.1109/ACC.2008.4587209

Filename

4587209