Active Learning with Disagreement Graphs
Corinna Cortes, Giulia Desalvo, Mehryar Mohri, Ningshan Zhang, Claudio Gentile
Abstract
We present two novel enhancements of an online importance-weighted active learning algorithm IWAL, using the properties of disagreements among hypotheses. The first enhancement, IWALD, prunes the hypothesis set with a more aggressive strategy based on the disagreement graph. We show that IWAL-D improves the generalization performance and the label complexity of the original IWAL, and quantify the improvement in terms of the disagreement graph coefficient. The second enhancement, IZOOM, further improves IWAL-D by adaptively zooming into the current version space and thus reducing the best-in-class error. We show that IZOOM admits favorable theoretical guarantees with the changing hypothesis set. We report experimental results on multiple datasets and demonstrate that the proposed algorithms achieve better test performances than IWAL given the same amount of labeling budget.
BibTeX
@InProceedings{pmlr-v97-cortes19b,
title = {Active Learning with Disagreement Graphs},
author = {Cortes, Corinna and Desalvo, Giulia and Mohri, Mehryar and Zhang, Ningshan and Gentile, Claudio},
booktitle = {Proceedings of the 36th International Conference on Machine Learning},
pages = {1379--1387},
year = {2019},
editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan},
volume = {97},
series = {Proceedings of Machine Learning Research},
month = {09--15 Jun},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v97/cortes19b/cortes19b.pdf},
url = {https://proceedings.mlr.press/v97/cortes19b.html},
abstract = {We present two novel enhancements of an online importance-weighted active learning algorithm IWAL, using the properties of disagreements among hypotheses. The first enhancement, IWALD, prunes the hypothesis set with a more aggressive strategy based on the disagreement graph. We show that IWAL-D improves the generalization performance and the label complexity of the original IWAL, and quantify the improvement in terms of the disagreement graph coefficient. The second enhancement, IZOOM, further improves IWAL-D by adaptively zooming into the current version space and thus reducing the best-in-class error. We show that IZOOM admits favorable theoretical guarantees with the changing hypothesis set. We report experimental results on multiple datasets and demonstrate that the proposed algorithms achieve better test performances than IWAL given the same amount of labeling budget.}
}