Blind Exploration and Exploitation of Stochastic Experts

TL;DR

This paper introduces blind exploration and exploitation algorithms for identifying reliable stochastic experts in multi-armed bandit problems, using unsupervised measures of competence to enable joint expert consultation without true state feedback.

Contribution

It proposes an empirically realizable competence measure that preserves true competence orderings and facilitates unsupervised expert aggregation and decision-making.

Findings

BEE algorithms perform comparably to supervised methods in experiments.

The proposed measure effectively ranks expert reliability without supervision.

Blind algorithms exhibit manageable regret growth similar to supervised counterparts.

Abstract

We present blind exploration and exploitation (BEE) algorithms for identifying the most reliable stochastic expert based on formulations that employ posterior sampling, upper-confidence bounds, empirical Kullback-Leibler divergence, and minmax methods for the stochastic multi-armed bandit problem. Joint sampling and consultation of experts whose opinions depend on the hidden and random state of the world becomes challenging in the unsupervised, or blind, framework as feedback from the true state is not available. We propose an empirically realizable measure of expert competence that can be inferred instantaneously using only the opinions of other experts. This measure preserves the ordering of true competences and thus enables joint sampling and consultation of stochastic experts based on their opinions on dynamically changing tasks. Statistics derived from the proposed measure is instantaneously available allowing both blind exploration-exploitation and unsupervised opinion aggregation. We discuss how the lack of supervision affects the asymptotic regret of BEE architectures that rely on UCB1, KL-UCB, MOSS, IMED, and Thompson sampling. We demonstrate the performance of different BEE algorithms empirically and compare them to their standard, or supervised, counterparts.