Similarity Metric Learning for a Variable-Kernel Classifier

  title={Similarity Metric Learning for a Variable-Kernel Classifier},
  author={David G. Lowe},
  journal={Neural Computation},
Nearest-neighbor interpolation algorithms have many useful properties for applications to learning, but they often exhibit poor generalization. In this paper, it is shown that much better generalization can be obtained by using a variable interpolation kernel in combination with conjugate gradient optimization of the similarity metric and kernel size. The resulting method is called variable-kernel similarity metric (VSM) learning. It has been tested on several standard classification data sets… CONTINUE READING
Highly Influential
This paper has highly influenced 16 other papers. REVIEW HIGHLY INFLUENTIAL CITATIONS
Highly Cited
This paper has 342 citations. REVIEW CITATIONS


Publications citing this paper.
Showing 1-10 of 183 extracted citations

342 Citations

Citations per Year
Semantic Scholar estimates that this publication has 342 citations based on the available data.

See our FAQ for additional information.


Publications referenced by this paper.
Showing 1-10 of 22 references

Dynamic error propagation networks

  • J.
  • Ph.D. thesis, Cambridge University Engineering…
  • 1989
Highly Influential
5 Excerpts

Improving the performance of radial basis function networks by learning center locations

  • D. H. Wolpert
  • Aduances in Neural Information Processing Systems
  • 1992

NN concepts and techniques

  • B. V. Dasarathy
  • Nearest Neighbour ( N N ) Norms: N N Pattern…
  • 1991
2 Excerpts

A theory of networks for approximation and learning

  • T. Poggio, F. Girosi
  • Report AI-1140, MIT Artificial Intelligence…
  • 1989

Similar Papers

Loading similar papers…