English

Using Stochastic Encoders to Discover Structure in Data

Neural and Evolutionary Computing 2007-05-23 v1 Computer Vision and Pattern Recognition

Abstract

In this paper a stochastic generalisation of the standard Linde-Buzo-Gray (LBG) approach to vector quantiser (VQ) design is presented, in which the encoder is implemented as the sampling of a vector of code indices from a probability distribution derived from the input vector, and the decoder is implemented as a superposition of reconstruction vectors. This stochastic VQ (SVQ) is optimised using a minimum mean Euclidean reconstruction distortion criterion, as in the LBG case. Numerical simulations are used to demonstrate how this leads to self-organisation of the SVQ, where different stochastically sampled code indices become associated with different input subspaces.

Cite

@article{arxiv.cs/0408049,
  title  = {Using Stochastic Encoders to Discover Structure in Data},
  author = {Stephen Luttrell},
  journal= {arXiv preprint arXiv:cs/0408049},
  year   = {2007}
}

Comments

18 pages, 9 figures. Full version of a short paper that was published in the Digest of the 5th IMA International Conference on Mathematics in Signal Processing, 18-20 December 2000, Warwick University, UK