Noisy desired signal in transient detection using neural networks

Jose C. Principe; Abir Zahalka

doi:10.1117/12.162722

5 November 1993 Noisy desired signal in transient detection using neural networks

Jose C. Principe, Abir Zahalka

Proceedings Volume 2036, Chaos in Biology and Medicine; (1993) https://doi.org/10.1117/12.162722
Event: SPIE's 1993 International Symposium on Optics, Imaging, and Instrumentation, 1993, San Diego, CA, United States

Abstract

In this paper we show the effect that several desired signals have on the performance of a neural network dynamic classifier for transient detection. We compare performances of the same neural network trained with the conventional 1/0 desired signal, a prediction framework and a desired signal composed of noise during the background. This last choice is the one that works best. We show that in terms of statistical decision theory this choice of desired signal should work as well as the optimal a posteriori detector. We provide an explanation why the noise during the background works for transient detection. Finally we comment on the implications of this choice of desired signal for learning in biological networks.

Citation Download Citation

Jose C. Principe and Abir Zahalka "Noisy desired signal in transient detection using neural networks", Proc. SPIE 2036, Chaos in Biology and Medicine, (5 November 1993); https://doi.org/10.1117/12.162722

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available