Generalized Risk-Aversion in Stochastic Multi-Armed Bandits

Abstract

We consider the problem of minimizing the regret in stochastic multi-armed bandit, when the measure of goodness of an arm is not the mean return, but some general function of the mean and the variance.We characterize the conditions under which learning is possible and present examples for which no natural algorithm can achieve sublinear regret.

References in corpus (4)

• The KL-UCB Algorithm for Bounded Stochastic Bandits and Beyond
• A Finite-Time Analysis of Multi-armed Bandits Problems with Kullback-Leibler Divergences
• Risk-sensitive Markov control processes
• Exploration vs Exploitation vs Safety: Risk-averse Multi-Armed Bandits

Cited by in corpus (7)

• Risk-Averse Multi-Armed Bandit Problems under Mean-Variance Measure
• Risk-Sensitive Reinforcement Learning: Near-Optimal Risk-Sample Tradeoff in Regret
• The Fragility of Optimized Bandit Algorithms
• Functional Sequential Treatment Allocation with Covariates
• The Survival Bandit Problem
• Decision Variance in Online Learning
• Rarely-switching linear bandits: optimization of causal effects for the real world