Hierarchical Label Queries with Data-Dependent Partitions
Abstract
Given a joint distribution P_X, Y over a space \Xcal and a label set \Ycal=\braces0, 1, we consider the problem of recovering the labels of an unlabeled sample with as few label queries as possible. The recovered labels can be passed to a passive learner, thus turning the procedure into an active learning approach. We analyze a family of labeling procedures based on a hierarchical clustering of the data. While such labeling procedures have been studied in the past, we provide a new parametrization of P_X, Y that captures their behavior in general low-noise settings, and which accounts for data-dependent clustering, thus providing new theoretical underpinning to practically used tools.
Cite
Text
Kpotufe et al. "Hierarchical Label Queries with Data-Dependent Partitions." Annual Conference on Computational Learning Theory, 2015.Markdown
[Kpotufe et al. "Hierarchical Label Queries with Data-Dependent Partitions." Annual Conference on Computational Learning Theory, 2015.](https://mlanthology.org/colt/2015/kpotufe2015colt-hierarchical/)BibTeX
@inproceedings{kpotufe2015colt-hierarchical,
title = {{Hierarchical Label Queries with Data-Dependent Partitions}},
author = {Kpotufe, Samory and Urner, Ruth and Ben-David, Shai},
booktitle = {Annual Conference on Computational Learning Theory},
year = {2015},
pages = {1176-1189},
url = {https://mlanthology.org/colt/2015/kpotufe2015colt-hierarchical/}
}