Spectral Risk-Based Learning Using Unbounded Losses
Abstract
In this work, we consider the setting of learning problems under a wide class of spectral risk (or "L-risk") functions, where a Lipschitz-continuous spectral density is used to flexibly assign weight to extreme loss values. We obtain excess risk guarantees for a derivative-free learning procedure under unbounded heavy-tailed loss distributions, and propose a computationally efficient implementation which empirically outperforms traditional risk minimizers in terms of balancing spectral risk and misclassification error.
Cite
Text
Holland and Mehdi Haress. "Spectral Risk-Based Learning Using Unbounded Losses." Artificial Intelligence and Statistics, 2022.Markdown
[Holland and Mehdi Haress. "Spectral Risk-Based Learning Using Unbounded Losses." Artificial Intelligence and Statistics, 2022.](https://mlanthology.org/aistats/2022/holland2022aistats-spectral/)BibTeX
@inproceedings{holland2022aistats-spectral,
title = {{Spectral Risk-Based Learning Using Unbounded Losses}},
author = {Holland, Matthew J. and Mehdi Haress, El},
booktitle = {Artificial Intelligence and Statistics},
year = {2022},
pages = {1871-1886},
volume = {151},
url = {https://mlanthology.org/aistats/2022/holland2022aistats-spectral/}
}