• Kod.IJCNN2000


Prof. Daniel E. Koditschek

Dan’s Home

Kod*lab Menu



Internal Links (Login Required)



<< User’s Publications << Kod*lab Publications

Piecewise Linear Homeomorphisms: the Scalar Case

IEEE-INNS-ENNS 2000 International Joint Conference on Neural Networks, Como, Italy, 2000

Richard E. Groff, Daniel E. Koditschek, and Pramod P. Khargonekar
University of Michigan
Full PDF | Penn ScholarlyCommons | IEEExplore

Abstract
      The class of piecewise linear homeomorphisms (PLH) provides a convenient functional representation for many applications wherein an approximation to data is required that is invertible in closed form. In this paper we introduce the graph intersection (GI) algorithm for “learning” piecewise linear scalar functions in two settings: “approximation,” where an “oracle” outputs accurate functional values in response to input queries; and “estimation,” where only a fixed discrete data base of input-output pairs is available. We provide a local convergence result for the approximation version of the GI algorithm as well as a study of its numerical performance in the estimation setting. We conclude that PLH offers accuracy closed to that of a neural net while requiring, via our GI algorithm, far shorter training time and preserving desired invariant properties unlike any other presently popular basis family.
BibTeX entry
@inproceedings{Groff-IJCNN-2000
  author       = "Richard E. Groff, Daniel E. Koditschek, Pramod P. Khargonekar"
  title        = "Piecewise Linear Homeomorphisms: the Scalar Case"
  publisher    = "IEEE"
  booktitle    = "Proceedings of IEEE-INNS-ENNS International Joint Conference on Neural Networks"
  volume       = "3"
  pages        = "259-264"
  year         = "2000"
}

Copyright Kodlab, 2017