Regret minimization in online Bayesian persuasion: Handling adversarial receiver's types under full and partial feedback models

Castiglioni, Matteo; Celli, Andrea; Marchesi, Alberto; Gatti, Nicola

doi:10.1016/j.artint.2022.103821

Cited by 4 publications

(1 citation statement)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work also relates to a recent trend of works on Bayesian persuasion with learning: e.g., Castiglioni et al (2020Castiglioni et al ( , 2021; Zu et al (2021); Feng et al (2022b); Wu et al (2022). The learner in these works is the sender, who repeatedly interacts with the receiver, learns additional information over time, and adjusts the signaling scheme accordingly.…”

Section: Related Workmentioning

confidence: 93%

Persuading a Behavioral Agent: Approximately Best Responding and Learning

Lin¹

2023

Preprint

View full text Add to dashboard Cite

The classic Bayesian persuasion model assumes a Bayesian and best-responding receiver. We study a relaxation of the Bayesian persuasion model where the receiver can approximately best respond to the sender's signaling scheme. We show that, under natural assumptions, (1) the sender can find a signaling scheme that guarantees itself an expected utility almost as good as its optimal utility in the classic model, no matter what approximately best-responding strategy the receiver uses; (2) on the other hand, there is no signaling scheme that gives the sender much more utility than its optimal utility in the classic model, even if the receiver uses the approximately best-responding strategy that is best for the sender. Together, (1) and ( 2) imply that the approximately best-responding behavior of the receiver does not affect the sender's maximal achievable utility a lot in the Bayesian persuasion problem. The proofs of both results rely on the idea of robustification of a Bayesian persuasion scheme: given a pair of the sender's signaling scheme and the receiver's strategy, we can construct another signaling scheme such that the receiver prefers to use that strategy in the new scheme more than in the original scheme, and the two schemes give the sender similar utilities. As an application of our main result (1), we show that, in a repeated Bayesian persuasion model where the receiver learns to respond to the sender by some algorithms, the sender can do almost as well as in the classic model. Interestingly, unlike (2), with a learning receiver the sender can sometimes do much better than in the classic model.

show abstract

Section: Related Workmentioning

confidence: 93%

Persuading a Behavioral Agent: Approximately Best Responding and Learning

Lin¹

2023

Preprint

View full text Add to dashboard Cite

show abstract

Reducing the Gap Between Theory and Applications in Algorithmic Bayesian Persuasion

Castiglioni

2024

SpringerBriefs in Applied Sciences and Technology

View full text Add to dashboard Cite

This work focuses on the following question: is it possible to influence the behavior of self-interested agents through the strategic provision of information? This ‘sweet talk’ is ubiquitous among all sorts of economics and non-economics activities. In this work, we model these multi-agent systems as games between an informed sender and one or multiple receivers. We study the computational problem faced by an informed sender that wants to use his information advantage to influence rational receivers with the partial disclosure of information. In particular, the sender faces an information structure design problem that amounts to deciding ‘who gets to know what’. Bayesian persuasion provides a formal framework to model these settings as asymmetric-information games. In recent years, much attention has been given to Bayesian persuasion in the economics and artificial intelligence communities due also to the applicability of this framework to a large class of scenarios like online advertising, voting, traffic routing, recommendation systems, security, and product marketing. However, there is still a large gap between the theoretical study of information in games and its applications in real-world scenarios. This work contributes to close this gap along two directions. First, we study the persuasion problem in real-world scenarios, focusing on voting, routing, and auctions. While the Bayesian persuasion framework can be applied to all these settings, the algorithmic problem of designing optimal information disclosure polices introduces computational challenges related to the specific problem under study. Our goal is to settle the complexity of computing optimal sender’s strategies, showing when an optimal strategy can be implemented efficiently. Then, we relax stringent assumptions that limit the applicability of the Bayesian persuasion framework in practice. In particular, the classical model assumes that the sender has perfect knowledge of the receiver’s utility. We remove this assumption initiating the study of an online version of the persuasion problem. This is the first step in designing adaptive information disclosure policies that deal with the uncertainty intrinsic in all real-world applications.

show abstract