Title :
Backpropagation for linearly-separable patterns: A detailed analysis
Author :
Rasconi, Paolo F. ; Gori, Marco ; Tesi, Albert0
Author_Institution :
Dipartimento di Sistemi e Inf., Florence Univ., Italy
Abstract :
A sufficient condition for learning without local minima in multilayered networks is proposed. A fundamental assumption on the network architecture is removed. It is proved that the conclusions drawn by M. Gori and A. Tesi (IEEE Trans. Pattern Anal. Mach. Intell., vol.14, no.1, pp.76-86, (1992)) also hold provided that the weight matrix associated with the hidden and output layer is pyramidal and has full rank. The analysis is carried out by using least mean squares (LMS)-threshold cost functions, which allow the identification of spurious and structural local minima
Keywords :
backpropagation; feedforward neural nets; matrix algebra; backpropagation; learning; least mean squares; local minima; multilayered networks; neural nets; sufficient condition; threshold cost functions; weight matrix; Algorithm design and analysis; Backpropagation algorithms; Cost function; Electronic mail; Interpolation; Joining processes; Neurons; Pattern analysis; Shape; Sufficient conditions;
Conference_Titel :
Neural Networks, 1993., IEEE International Conference on
Conference_Location :
San Francisco, CA
Print_ISBN :
0-7803-0999-5
DOI :
10.1109/ICNN.1993.298833
