The Convexity and Design of Composite Multiclass Losses
Abstract
We consider composite loss functions for multiclass prediction comprising a proper (i.e., Fisher-consistent) loss over probability distributions and an inverse link function. We establish conditions for their (strong) convexity and explore the implications. We also show how the separation of concerns afforded by using this composite representation allows for the design of families of losses with the same Bayes risk.
Cite
Text
Reid et al. "The Convexity and Design of Composite Multiclass Losses." International Conference on Machine Learning, 2012.Markdown
[Reid et al. "The Convexity and Design of Composite Multiclass Losses." International Conference on Machine Learning, 2012.](https://mlanthology.org/icml/2012/reid2012icml-convexity/)BibTeX
@inproceedings{reid2012icml-convexity,
title = {{The Convexity and Design of Composite Multiclass Losses}},
author = {Reid, Mark D. and Williamson, Robert C. and Sun, Peng},
booktitle = {International Conference on Machine Learning},
year = {2012},
url = {https://mlanthology.org/icml/2012/reid2012icml-convexity/}
}