Active Learning for Identifying Function Threshold Boundaries

Abstract

We present an efficient algorithm to actively select queries for learning the boundaries separating a function domain into regions where the func- tion is above and below a given threshold. We develop experiment selec- tion methods based on entropy, misclassification rate, variance, and their combinations, and show how they perform on a number of data sets. We then show how these algorithms are used to determine simultaneously valid 1 − α confidence intervals for seven cosmological parameters. Ex- perimentation shows that the algorithm reduces the computation neces- sary for the parameter estimation problem by an order of magnitude.

Cite

Text

Bryan et al. "Active Learning for Identifying Function Threshold Boundaries." Neural Information Processing Systems, 2005.

Markdown

[Bryan et al. "Active Learning for Identifying Function Threshold Boundaries." Neural Information Processing Systems, 2005.](https://mlanthology.org/neurips/2005/bryan2005neurips-active/)

BibTeX

@inproceedings{bryan2005neurips-active,
  title     = {{Active Learning for Identifying Function Threshold Boundaries}},
  author    = {Bryan, Brent and Nichol, Robert C. and Genovese, Christopher R and Schneider, Jeff and Miller, Christopher J. and Wasserman, Larry},
  booktitle = {Neural Information Processing Systems},
  year      = {2005},
  pages     = {163-170},
  url       = {https://mlanthology.org/neurips/2005/bryan2005neurips-active/}
}