Title :
Tolerantly linear separability of Boolean functions and its numbering
Author :
Junying, Zhang ; Jin, Xu
Author_Institution :
Lab. of Radar Signal Process., Xidian Univ., Xi´´an, China
Abstract :
Linear and non-linear separability of Boolean functions are difficult problems in feedforward neural networks. Only linear separability of Boolean functions with the number of variables n<7 had been discussed before. In this paper, a tolerantly linear-separable complexity of Boolean functions is defined to simplify discussions and optimize solutions. An approach of numbering Boolean functions whose tolerantly linear-separable complexity is c is presented, and the results for c=1 are presented
Keywords :
Boolean functions; computational complexity; feedforward neural nets; hypercube networks; Boolean functions; feedforward neural networks; numbering; tolerantly linear-separable complexity; Auditory system; Boolean functions; Equations; Feedforward neural networks; Hypercubes; Laboratories; Neural networks; Signal processing; Speech processing;
Conference_Titel :
Signal Processing, 1996., 3rd International Conference on
Conference_Location :
Beijing
Print_ISBN :
0-7803-2912-0
DOI :
10.1109/ICSIGP.1996.571129
