ON UNSUPERVISED NEAREST-NEIGHBOR REGRESSION AND ROBUST LOSS FUNCTIONS
Oliver Kramer
2012
Abstract
In many scientific disciplines structures in high-dimensional data have to be detected, e.g., in stellar spectra, in genome data, or in face recognition tasks. We present an approach to non-linear dimensionality reduction based on fitting nearest neighbor regression to the unsupervised regression framework for learning of lowdimensional manifolds. The problem of optimizing latent neighborhoods is difficult to solve, but the UNN formulation allows an efficient strategy of iteratively embedding latent points to fixed neighborhood topologies. The choice of an appropriate loss function is relevant, in particular for noisy, and high-dimensional data spaces. We extend unsupervised nearest neighbor (UNN) regression by the e-insensitive loss, which allows to ignore residuals under a threshold defined by e. In the experimental part of this paper we test the influence of e on the final data space reconstruction error, and present a visualization of UNN embeddings on test data sets.
DownloadPaper Citation
in Harvard Style
Kramer O. (2012). ON UNSUPERVISED NEAREST-NEIGHBOR REGRESSION AND ROBUST LOSS FUNCTIONS . In Proceedings of the 4th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART, ISBN 978-989-8425-95-9, pages 164-170. DOI: 10.5220/0003749301640170
in Bibtex Style
@conference{icaart12,
author={Oliver Kramer},
title={ON UNSUPERVISED NEAREST-NEIGHBOR REGRESSION AND ROBUST LOSS FUNCTIONS},
booktitle={Proceedings of the 4th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART,},
year={2012},
pages={164-170},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0003749301640170},
isbn={978-989-8425-95-9},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 4th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART,
TI - ON UNSUPERVISED NEAREST-NEIGHBOR REGRESSION AND ROBUST LOSS FUNCTIONS
SN - 978-989-8425-95-9
AU - Kramer O.
PY - 2012
SP - 164
EP - 170
DO - 10.5220/0003749301640170