Active Heteroscedastic Regression

Kamalika Chaudhuri, Prateek Jain, Nagarajan Natarajan
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:694-702, 2017.

Abstract

An active learner is given a model class $\Theta$, a large sample of unlabeled data drawn from an underlying distribution and access to a labeling oracle that can provide a label for any of the unlabeled instances. The goal of the learner is to find a model $\theta \in \Theta$ that fits the data to a given accuracy while making as few label queries to the oracle as possible. In this work, we consider a theoretical analysis of the label requirement of active learning for regression under a heteroscedastic noise model, where the noise depends on the instance. We provide bounds on the convergence rates of active and passive learning for heteroscedastic regression. Our results illustrate that just like in binary classification, some partial knowledge of the nature of the noise can lead to significant gains in the label requirement of active learning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-chaudhuri17a, title = {Active Heteroscedastic Regression}, author = {Kamalika Chaudhuri and Prateek Jain and Nagarajan Natarajan}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {694--702}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/chaudhuri17a/chaudhuri17a.pdf}, url = {https://proceedings.mlr.press/v70/chaudhuri17a.html}, abstract = {An active learner is given a model class $\Theta$, a large sample of unlabeled data drawn from an underlying distribution and access to a labeling oracle that can provide a label for any of the unlabeled instances. The goal of the learner is to find a model $\theta \in \Theta$ that fits the data to a given accuracy while making as few label queries to the oracle as possible. In this work, we consider a theoretical analysis of the label requirement of active learning for regression under a heteroscedastic noise model, where the noise depends on the instance. We provide bounds on the convergence rates of active and passive learning for heteroscedastic regression. Our results illustrate that just like in binary classification, some partial knowledge of the nature of the noise can lead to significant gains in the label requirement of active learning.} }
Endnote
%0 Conference Paper %T Active Heteroscedastic Regression %A Kamalika Chaudhuri %A Prateek Jain %A Nagarajan Natarajan %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-chaudhuri17a %I PMLR %P 694--702 %U https://proceedings.mlr.press/v70/chaudhuri17a.html %V 70 %X An active learner is given a model class $\Theta$, a large sample of unlabeled data drawn from an underlying distribution and access to a labeling oracle that can provide a label for any of the unlabeled instances. The goal of the learner is to find a model $\theta \in \Theta$ that fits the data to a given accuracy while making as few label queries to the oracle as possible. In this work, we consider a theoretical analysis of the label requirement of active learning for regression under a heteroscedastic noise model, where the noise depends on the instance. We provide bounds on the convergence rates of active and passive learning for heteroscedastic regression. Our results illustrate that just like in binary classification, some partial knowledge of the nature of the noise can lead to significant gains in the label requirement of active learning.
APA
Chaudhuri, K., Jain, P. & Natarajan, N.. (2017). Active Heteroscedastic Regression. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:694-702 Available from https://proceedings.mlr.press/v70/chaudhuri17a.html.

Related Material