Inverse Viterbi algorithm as learning procedure and application to optimization in the string instrument fingering problem

Author

Sayegh, Samir I. ; Tenorio, Manoel F.

Author_Institution

Dept. of Phys., Indiana-Purdue Univ., Fort Wayne, IN, USA

fYear

1988

fDate

24-27 July 1988

Firstpage

491

Abstract

The realization of a certain function is implemented as a minimization principle. The input is specified by activating a subnetwork of the network and the output by finding the best path connecting the first layer to the last layer of the active subnetwork. This formulation is very attractive because of the existence of fast parallel algorithms for the implementation of the forward phase and because of the simplicity and speed of the learning phase algorithm that the authors propose. A large class of functions can naturally be implemented within this framework. The particular application chosen is that of the optimization problem for stringed instrument fingering.<>

Keywords

humanities; learning systems; musical instruments; neural nets; optimisation; parallel algorithms; fast parallel algorithms; guitar; inverse Viterbi algorithm; learning procedure; minimization principle; optimization; stringed instrument fingering; subnetwork activation; Humanities; Learning systems; Neural networks; Optimization methods; Parallel algorithms;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 1988., IEEE International Conference on

Conference_Location

San Diego, CA, USA

Type

conf

DOI

10.1109/ICNN.1988.23964

Filename

23964