Stochastic Optimization for Spectral Risk Measures

arXiv2212.05149MaRDI QIDQ6420104

Author name not available (Why is that?)

Publication date: 9 December 2022

Abstract: Spectral risk objectives - also called $L$-risks - allow for learning systems to interpolate between optimizing average-case performance (as in empirical risk minimization) and worst-case performance on a task. We develop stochastic algorithms to optimize these quantities by characterizing their subdifferential and addressing challenges such as biasedness of subgradient estimates and non-smoothness of the objective. We show theoretically and experimentally that out-of-the-box approaches such as stochastic subgradient and dual averaging are hindered by bias and that our approach outperforms them.

Has companion code repository: https://github.com/ronakdm/lerm

This page was built for publication: Stochastic Optimization for Spectral Risk Measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6420104)