Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shao, Xiao, Wu, Guoqiang
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2502.18167
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910958390083584
author Shao, Xiao
Wu, Guoqiang
author_facet Shao, Xiao
Wu, Guoqiang
contents In multi-task learning (MTL) with each task involving graph-dependent data, existing generalization analyses yield a \emph{sub-optimal} risk bound of $O(\frac{1}{\sqrt{n}})$, where $n$ is the number of training samples of each task. However, to improve the risk bound is technically challenging, which is attributed to the lack of a foundational sharper concentration inequality for multi-graph dependent random variables. To fill up this gap, this paper proposes a new Bennett-type inequality, enabling the derivation of a sharper risk bound of $O(\frac{\log n}{n})$. Technically, building on the proposed Bennett-type inequality, we propose a new Talagrand-type inequality for the empirical process, and further develop a new analytical framework of the local fractional Rademacher complexity to enhance generalization analyses in MTL with multi-graph dependent data. Finally, we apply the theoretical advancements to applications such as Macro-AUC optimization, illustrating the superiority of our theoretical results over prior work, which is also verified by experimental results.
format Preprint
id arxiv_https___arxiv_org_abs_2502_18167
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sharper Risk Bound for Multi-Task Learning with Multi-Graph Dependent Data
Shao, Xiao
Wu, Guoqiang
Machine Learning
In multi-task learning (MTL) with each task involving graph-dependent data, existing generalization analyses yield a \emph{sub-optimal} risk bound of $O(\frac{1}{\sqrt{n}})$, where $n$ is the number of training samples of each task. However, to improve the risk bound is technically challenging, which is attributed to the lack of a foundational sharper concentration inequality for multi-graph dependent random variables. To fill up this gap, this paper proposes a new Bennett-type inequality, enabling the derivation of a sharper risk bound of $O(\frac{\log n}{n})$. Technically, building on the proposed Bennett-type inequality, we propose a new Talagrand-type inequality for the empirical process, and further develop a new analytical framework of the local fractional Rademacher complexity to enhance generalization analyses in MTL with multi-graph dependent data. Finally, we apply the theoretical advancements to applications such as Macro-AUC optimization, illustrating the superiority of our theoretical results over prior work, which is also verified by experimental results.
title Sharper Risk Bound for Multi-Task Learning with Multi-Graph Dependent Data
topic Machine Learning
url https://arxiv.org/abs/2502.18167