Meta-Learning for Repeated Bayesian Persuasion

TL;DR

This paper introduces Meta-Persuasion algorithms for repeated Bayesian persuasion, achieving sharper regret bounds by leveraging task similarity and demonstrating benefits through theoretical analysis and numerical experiments.

Contribution

It presents the first theoretical meta-learning algorithms for repeated Bayesian persuasion in both full-feedback and bandit settings, improving convergence rates.

Findings

Sharper regret bounds under task similarity

Algorithms recover single-game guarantees for arbitrary sequences

Numerical experiments show regret improvements and benefits of meta-learning

Abstract

Classical Bayesian persuasion studies how a sender influences receivers through carefully designed signaling policies within a single strategic interaction. In many real-world environments, such interactions are repeated across multiple games, creating opportunities to exploit structural similarity across tasks. In this work, we introduce Meta-Persuasion algorithms, establishing the first line of theoretical results for both full-feedback and bandit-feedback settings in the Online Bayesian Persuasion (OBP) and Markov Persuasion Process (MPP) frameworks. We show that our proposed meta-persuasion algorithms achieve provably sharper regret rates under natural notions of task similarity, improving upon the best-known convergence rates for both OBP and MPP. At the same time, they recover the standard single-game guarantees when the sequence of games is picked arbitrarily. Finally, we complement our theoretical analysis with numerical experiments that highlight our regret improvements and the benefits of meta-learning in repeated persuasion environments.