Timezone: »

 
Planning to Fairly Allocate: Probabilistic Fairness in the Restless Bandit Setting
Christine Herlihy · Aviva Prins · Aravind Srinivasan · John P Dickerson

@
in Responsible Decision Making in Dynamic Environments »
Restless and collapsing bandits are often used to model budget-constrained resource allocation in settings where receiving the resource increases the probability that an arm will transition to, or remain in, a desirable state. However, SOTA Whittle-index-based approaches to this planning problem either do not consider fairness among arms, or incentivize fairness without guaranteeing it. We introduce ProbFair, an algorithm which finds the best (reward-maximizing) policy that: (a) satisfies the budget constraint; and (b) enforces bounds $[\ell, u]$ on the probability of being pulled at each timestep. We evaluate our algorithm on a real-world application, where interventions support continuous positive airway pressure (CPAP) therapy adherence among patients, as well as on a broader class of synthetic transition matrices. ProbFair preserves utility while providing fairness guarantees.

Author Information

Christine Herlihy (University of Maryland, College Park)
Aviva Prins (University of Maryland, College Park)
Aravind Srinivasan (Amazon)
John P Dickerson (Arthur AI & Univ. of Maryland)

More from the Same Authors