Fahlman-Type Activation Functions Applied to Nonlinear PCA Networks Provide a Generalised Independent Component Analysis

Mark Girolami, Colin Fyfe. Fahlman-Type Activation Functions Applied to Nonlinear PCA Networks Provide a Generalised Independent Component Analysis. In George D. Smith, Nigel C. Steele, Rudolf F. Albrecht, editors, Proceedings of the International Conference on Artificial Neural Nets and Genetic Algorithms, ICANNGA 1997, Norwich, UK, 1997. pages 112-115, Springer, 1997. [doi]

Abstract

Abstract is missing.