Multi-armed linear bandits with latent biases

In a linear stochastic bandit model, each arm corresponds to a vector in Euclidean space, and the expected return observed at each time step is determined by an unknown linear function of the selected arm. This paper addresses the challenge of identifying the optimal arm in a linear stochastic bandi...

Full description

Saved in:
Bibliographic Details
Main Authors: Kang, Qiyu, Tay, Wee Peng, She, Rui, Wang, Sijie, Liu, Xiaoqian, Yang, Yuan-Rui
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2024
Subjects:
Online Access:https://hdl.handle.net/10356/175416
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-175416
record_format dspace
spelling sg-ntu-dr.10356-1754162024-04-26T15:55:30Z Multi-armed linear bandits with latent biases Kang, Qiyu Tay, Wee Peng She, Rui Wang, Sijie Liu, Xiaoqian Yang, Yuan-Rui School of Electrical and Electronic Engineering Computer and Information Science Linear bandit Multi-armed bandit In a linear stochastic bandit model, each arm corresponds to a vector in Euclidean space, and the expected return observed at each time step is determined by an unknown linear function of the selected arm. This paper addresses the challenge of identifying the optimal arm in a linear stochastic bandit model, where latent biases corrupt each arm's expected reward. Unlike traditional linear bandit problems, where the observed return directly represents the reward, this paper considers a scenario where the unbiased reward at each time step remains unobservable. This model is particularly relevant in situations where the observed return is influenced by latent biases that need to be carefully excluded from the learning model. For example, in recommendation systems designed to prevent racially discriminatory suggestions, it is crucial to ensure that the users' race does not influence the system. However, the observed return, such as click-through rates, may have already been influenced by racial attributes. In the case where there are finitely many arms, we develop a strategy to achieve O(|D|log⁡n) regret, where |D| is the number of arms and n is the number of time steps. In the case where each arm is chosen from an infinite compact set, our strategy achieves O(n2/3(log⁡n)1/2) regret. Experiments verify the efficiency of our strategy. Ministry of Education (MOE) Submitted/Accepted version This research is supported by the Singapore Ministry of Education Academic Research Fund Tier 2 grant MOE-T2EP20220-0002. 2024-04-23T05:32:23Z 2024-04-23T05:32:23Z 2024 Journal Article Kang, Q., Tay, W. P., She, R., Wang, S., Liu, X. & Yang, Y. (2024). Multi-armed linear bandits with latent biases. Information Sciences, 660, 120103-. https://dx.doi.org/10.1016/j.ins.2024.120103 0020-0255 https://hdl.handle.net/10356/175416 10.1016/j.ins.2024.120103 2-s2.0-85183462419 660 120103 en MOE-T2EP20220-0002 Information Sciences © 2024 Elsevier Inc. All rights reserved. This article may be downloaded for personal use only. Any other use requires prior permission of the copyright holder. The Version of Record is available online at http://doi.org/10.1016/j.ins.2024.120103. application/pdf
institution Nanyang Technological University
building NTU Library
continent Asia
country Singapore
Singapore
content_provider NTU Library
collection DR-NTU
language English
topic Computer and Information Science
Linear bandit
Multi-armed bandit
spellingShingle Computer and Information Science
Linear bandit
Multi-armed bandit
Kang, Qiyu
Tay, Wee Peng
She, Rui
Wang, Sijie
Liu, Xiaoqian
Yang, Yuan-Rui
Multi-armed linear bandits with latent biases
description In a linear stochastic bandit model, each arm corresponds to a vector in Euclidean space, and the expected return observed at each time step is determined by an unknown linear function of the selected arm. This paper addresses the challenge of identifying the optimal arm in a linear stochastic bandit model, where latent biases corrupt each arm's expected reward. Unlike traditional linear bandit problems, where the observed return directly represents the reward, this paper considers a scenario where the unbiased reward at each time step remains unobservable. This model is particularly relevant in situations where the observed return is influenced by latent biases that need to be carefully excluded from the learning model. For example, in recommendation systems designed to prevent racially discriminatory suggestions, it is crucial to ensure that the users' race does not influence the system. However, the observed return, such as click-through rates, may have already been influenced by racial attributes. In the case where there are finitely many arms, we develop a strategy to achieve O(|D|log⁡n) regret, where |D| is the number of arms and n is the number of time steps. In the case where each arm is chosen from an infinite compact set, our strategy achieves O(n2/3(log⁡n)1/2) regret. Experiments verify the efficiency of our strategy.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Kang, Qiyu
Tay, Wee Peng
She, Rui
Wang, Sijie
Liu, Xiaoqian
Yang, Yuan-Rui
format Article
author Kang, Qiyu
Tay, Wee Peng
She, Rui
Wang, Sijie
Liu, Xiaoqian
Yang, Yuan-Rui
author_sort Kang, Qiyu
title Multi-armed linear bandits with latent biases
title_short Multi-armed linear bandits with latent biases
title_full Multi-armed linear bandits with latent biases
title_fullStr Multi-armed linear bandits with latent biases
title_full_unstemmed Multi-armed linear bandits with latent biases
title_sort multi-armed linear bandits with latent biases
publishDate 2024
url https://hdl.handle.net/10356/175416
_version_ 1814047090047713280