Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Di Gennaro, Federico, Eldowa, Khaled, Cesa-Bianchi, Nicolò
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918288718561280
author Di Gennaro, Federico
Eldowa, Khaled
Cesa-Bianchi, Nicolò
author_facet Di Gennaro, Federico
Eldowa, Khaled
Cesa-Bianchi, Nicolò
contents In contrast to the classic formulation of partial monitoring, linear partial monitoring can model infinite outcome spaces, while imposing a linear structure on both the losses and the observations. This setting can be viewed as a generalization of linear bandits where loss and feedback are decoupled in a flexible manner. In this work, we address a nonstochastic (adversarial), finite-actions version of the problem through a simple instance of the exploration-by-optimization method that is amenable to efficient implementation. We derive regret bounds that depend on the game structure in a more transparent manner than previous theoretical guarantees for this paradigm. Our bounds feature instance-specific quantities that reflect the degree of alignment between observations and losses, and resemble known guarantees in the stochastic setting. Notably, they achieve the standard $\sqrt{T}$ rate in easy (locally observable) games and $T^{2/3}$ in hard (globally observable) games, where $T$ is the time horizon. We instantiate these bounds in a selection of old and new partial information settings subsumed by this model, and illustrate that the achieved dependence on the game structure can be tight in interesting cases.
format Preprint
id arxiv_https___arxiv_org_abs_2510_19158
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring
Di Gennaro, Federico
Eldowa, Khaled
Cesa-Bianchi, Nicolò
Machine Learning
In contrast to the classic formulation of partial monitoring, linear partial monitoring can model infinite outcome spaces, while imposing a linear structure on both the losses and the observations. This setting can be viewed as a generalization of linear bandits where loss and feedback are decoupled in a flexible manner. In this work, we address a nonstochastic (adversarial), finite-actions version of the problem through a simple instance of the exploration-by-optimization method that is amenable to efficient implementation. We derive regret bounds that depend on the game structure in a more transparent manner than previous theoretical guarantees for this paradigm. Our bounds feature instance-specific quantities that reflect the degree of alignment between observations and losses, and resemble known guarantees in the stochastic setting. Notably, they achieve the standard $\sqrt{T}$ rate in easy (locally observable) games and $T^{2/3}$ in hard (globally observable) games, where $T$ is the time horizon. We instantiate these bounds in a selection of old and new partial information settings subsumed by this model, and illustrate that the achieved dependence on the game structure can be tight in interesting cases.
title Instance-Dependent Regret Bounds for Nonstochastic Linear Partial Monitoring
topic Machine Learning
url https://arxiv.org/abs/2510.19158