Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Beyond Static Bias: Adaptive Multi-Fidelity Bandits with Improving Proxies

About

As an extension of the classical multi-armed bandit problem, multi-fidelity multi-armed bandits (MF-MAB) enable individual arms to be evaluated using diverse feedback sources that vary in both cost and accuracy. Prior stochastic models typically assume fixed low-to-high fidelity discrepancies, whereas modern proxy sources, such as learning-based simulators and Large Language Models (LLMs), can be improved using additional calibration. We investigate adaptive MF-MAB with improving proxy sources, and focus on the canonical two-fidelity case in which the low-fidelity source becomes more informative with repeated use. To capture this dynamic, we introduce a selected-average mismatch bound that converts dynamic low-fidelity observations into improvement-aware confidence bounds for the high-fidelity target. We propose the Threshold-Based Adaptive Continuation Companion (TACC), an optimistic algorithm that uses a bounded continuation rule to decide when low-fidelity sampling remains cost-effective and when to escalate. We prove an instance-dependent regret bound showing that, for detected intermediate arms, adaptive continuation replaces logarithmic high-fidelity confirmation with bounded low-fidelity continuation. Experiments on synthetic bandits and an LLM-as-a-judge policy-evaluation task examine when continuation improves cost-weighted regret.

Muyun Lu, Haoyang Hong, Huazheng Wang, Ying Lin• 2026

Related benchmarks

TaskDatasetResultRank
Multi-fidelity Multi-armed BanditNLI shared evaluation pool
Mean Cost-Weighted Pseudo-Regret3.28e+3
18
Bandit Regret MinimizationVanishing-mismatch experiment
Mean Cost-Weighted Pseudo-Regret75.7
15
Bandit Regret MinimizationCheckpoint-based (evaluation)--
5
Multi-fidelity bandit optimizationLLM-as-a-judge residual-mismatch Λ=128000 (test)
Mean Cost-Weighted Pseudo-Regret4.02e+3
4
Cost-weighted pseudo-regret minimizationResidual-mismatch experiment final-budget--
4
Showing 5 of 5 rows

Other info

Follow for update