A taxonomy of neural network pattern classifiers is presented which includes four major groupings. Global discriminant classifiers use sigmoid or polynoniial computing elements that have “high” non-zero outputs over most of their input space. Local discriminant classifiers use Gaussian or other localized computing elements that have Lbhigli” non-zero outputs over only a small localized region o...

This I)aper presents an implemented multi-tal)e twolevel model capable of describing Semitie non-linear morphology. The computational fl'arnework behind the ettrrcnt work is motivated by [Kay 1987]; the fimnalism presented here is an extension to the formalism reported by [Puhnan art(1 Hepl)le. 1993]. The objectives of the current work are: to stay as close as possible, in spirit, to standard t...