Coping with Label Shift via Distributionally Robust Optimisation

arXiv2010.12230MaRDI QIDQ6352058

Author name not available (Why is that?)

Publication date: 23 October 2020

Abstract: The label shift problem refers to the supervised learning setting where the train and test label distributions do not match. Existing work addressing label shift usually assumes access to an emph{unlabelled} test sample. This sample may be used to estimate the test label distribution, and to then train a suitably re-weighted classifier. While approaches using this idea have proven effective, their scope is limited as it is not always feasible to access the target domain; further, they require repeated retraining if the model is to be deployed in emph{multiple} test environments. Can one instead learn a emph{single} classifier that is robust to arbitrary label shifts from a broad family? In this paper, we answer this question by proposing a model that minimises an objective based on distributionally robust optimisation (DRO). We then design and analyse a gradient descent-proximal mirror ascent algorithm tailored for large-scale problems to optimise the proposed objective. %, and establish its convergence. Finally, through experiments on CIFAR-100 and ImageNet, we show that our technique can significantly improve performance over a number of baselines in settings where label shift is present.

Has companion code repository: https://github.com/ShahryarBQ/DRO

This page was built for publication: Coping with Label Shift via Distributionally Robust Optimisation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6352058)