Title :
Perceptrons, PP, and the polynomial hierarchy
Author_Institution :
Dept. of Comput. Sci., Yale Univ., New Haven, CT, USA
Abstract :
The author constructs a predicate that is computable by a perceptron with linear size, order one, and exponential weights, but which cannot be computed by any perceptron having subexponential size, subpolynomial order, and subexponential weights. A consequence is that there is an oracle relative to which PNP is not contained in PP
Keywords :
computational complexity; neural chips; threshold logic; PP; exponential weights; linear size weights; oracle; order one weights; perceptron; polynomial hierarchy; predicate; Circuits; Complexity theory; Computational modeling; Computer science; Polynomials; Turing machines; Upper bound;
Conference_Titel :
Structure in Complexity Theory Conference, 1992., Proceedings of the Seventh Annual
Conference_Location :
Boston, MA
Print_ISBN :
0-8186-2955-X
DOI :
10.1109/SCT.1992.215376
