Exploration via Feature Perturbation in Contextual Bandits

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yi, Seouh-won, Oh, Min-hwan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908608744128512
author Yi, Seouh-won
Oh, Min-hwan
author_facet Yi, Seouh-won
Oh, Min-hwan
contents We propose feature perturbation, a simple yet effective exploration strategy for contextual bandits that injects randomness directly into feature inputs, instead of randomizing unknown parameters or adding noise to rewards. Remarkably, this algorithm achieves $\tilde{\mathcal{O}}(d\sqrt{T})$ worst-case regret bound for generalized linear contextual bandits, while avoiding the $\tilde{\mathcal{O}}(d^{3/2}\sqrt{T})$ regret typical of existing randomized bandit algorithms. Because our algorithm eschews parameter sampling, it is both computationally efficient and naturally extends to non-parametric or neural network models. We verify these advantages through empirical evaluations, demonstrating that feature perturbation not only surpasses existing methods but also unifies strong practical performance with the near-optimal regret guarantees.
format Preprint
id arxiv_https___arxiv_org_abs_2510_17390
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exploration via Feature Perturbation in Contextual Bandits
Yi, Seouh-won
Oh, Min-hwan
Machine Learning
We propose feature perturbation, a simple yet effective exploration strategy for contextual bandits that injects randomness directly into feature inputs, instead of randomizing unknown parameters or adding noise to rewards. Remarkably, this algorithm achieves $\tilde{\mathcal{O}}(d\sqrt{T})$ worst-case regret bound for generalized linear contextual bandits, while avoiding the $\tilde{\mathcal{O}}(d^{3/2}\sqrt{T})$ regret typical of existing randomized bandit algorithms. Because our algorithm eschews parameter sampling, it is both computationally efficient and naturally extends to non-parametric or neural network models. We verify these advantages through empirical evaluations, demonstrating that feature perturbation not only surpasses existing methods but also unifies strong practical performance with the near-optimal regret guarantees.
title Exploration via Feature Perturbation in Contextual Bandits
topic Machine Learning
url https://arxiv.org/abs/2510.17390