Unlearning Offline Stochastic Multi-armed Bandits

Abstract

arXiv:2605.00638v1 Announce Type: new Abstract: Machine unlearning aims to unlearn data points from a learned model, offering a principled way to process data-deletion requests and mitigate privacy risks without full retraining. Prior work has mainly studied unsupervised / supervised machine unlearning, leaving unlearning for sequential decision-making systems far less understood. We initiate the first study of a foundational sequential decision-making problem: offline stochastic multi-armed bandits (MAB). We formalize the privacy constraint for offline MAB and measure utility by the post-unlearning decision quality. We conduct a systematic study of both single- and multi-source unlearning scenarios under two data-generation models, the fixed-sample model and the distribution model. For these settings, our algorithmic design is built on two canonical base algorithms: Gaussian mechanism and rollback, and we propose adaptive algorithms that switch between them according to the data regi

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyye2026unlearning

Related papers

• Online Learning For Cooperative Multi-player Multi-armed Bandits (2021)5.24
• Is Offline Decision Making Possible With Only Few Samples? Reliable Decisions In Data-starved Bandits Via Trust Region Enhancement (2024)0.00
• Multi-agent Bandit Learning Through Heterogeneous Action Erasure Channels (2023)0.00
• Non-stationary Latent Auto-regressive Bandits (2024)0.00
• Reinforcement Unlearning (2023)4.52
• Learning For Bandits Under Action Erasures (2024)0.00
• Unified Framework Of Distributional Regret In Multi-armed Bandits And Reinforcement Learning (2026)0.00
• A Frequency-domain Analysis Of The Multi-armed Bandit Problem: A New Perspective On The Exploration-exploitation Trade-off (2025)0.00