Reinforcing RCTs with Multiple Priors while Learning about External Validity

TL;DR

This paper presents a Bayesian framework that integrates multiple prior sources into sequential experiments, enhancing robustness and performance in learning and decision-making under uncertainty.

Contribution

It introduces a multi-prior Bayesian approach for experimental design that effectively combines diverse prior information and evaluates policies on key decision criteria.

Findings

Framework is robust to invalid external priors

Achieves strong finite sample performance

Improves decision accuracy and reward optimization

Abstract

This paper introduces a framework for incorporating prior information into the design of sequential experiments. These sources may include past experiments, expert opinions, or the experimenter's intuition. We model the problem using a multi-prior Bayesian approach, mapping each source to a Bayesian model and aggregating them based on posterior probabilities. Policies are evaluated on three criteria: learning the parameters of payoff distributions, the probability of choosing the wrong treatment, and average rewards. Our framework demonstrates several desirable properties, including robustness to sources lacking external validity, while maintaining strong finite sample performance.