English

Multi-Label Learning with Stronger Consistency Guarantees

Machine Learning 2024-07-19 v1 Machine Learning

Abstract

We present a detailed study of surrogate losses and algorithms for multi-label learning, supported by HH-consistency bounds. We first show that, for the simplest form of multi-label loss (the popular Hamming loss), the well-known consistent binary relevance surrogate suffers from a sub-optimal dependency on the number of labels in terms of HH-consistency bounds, when using smooth losses such as logistic losses. Furthermore, this loss function fails to account for label correlations. To address these drawbacks, we introduce a novel surrogate loss, multi-label logistic loss, that accounts for label correlations and benefits from label-independent HH-consistency bounds. We then broaden our analysis to cover a more extensive family of multi-label losses, including all common ones and a new extension defined based on linear-fractional functions with respect to the confusion matrix. We also extend our multi-label logistic losses to more comprehensive multi-label comp-sum losses, adapting comp-sum losses from standard classification to the multi-label learning. We prove that this family of surrogate losses benefits from HH-consistency bounds, and thus Bayes-consistency, across any general multi-label loss. Our work thus proposes a unified surrogate loss framework benefiting from strong consistency guarantees for any multi-label loss, significantly expanding upon previous work which only established Bayes-consistency and for specific loss functions. Additionally, we adapt constrained losses from standard classification to multi-label constrained losses in a similar way, which also benefit from HH-consistency bounds and thus Bayes-consistency for any multi-label loss. We further describe efficient gradient computation algorithms for minimizing the multi-label logistic loss.

Keywords

Cite

@article{arxiv.2407.13746,
  title  = {Multi-Label Learning with Stronger Consistency Guarantees},
  author = {Anqi Mao and Mehryar Mohri and Yutao Zhong},
  journal= {arXiv preprint arXiv:2407.13746},
  year   = {2024}
}
R2 v1 2026-06-28T17:46:24.099Z