Bayesian persuasion and information design: perspectives and open issues

Kamenica, Emir; Kim, Kyungmin; Zapechelnyuk, Andriy

doi:10.1007/s00199-021-01383-4

Cited by 7 publications

(8 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This formulation of the problem is equivalent to a so-called optimal Bayesian persuasion problem intorduced in Kamenica and Gentzkow (2011) (see also Rayo and Segal (2010), Dworczak and Kolotilin (2019), and Bergemann and Morris (2019), Kamenica et al (2021) for excellent overviews).…”

Section: E2 Generalized Optimal Autoencoder Formulationmentioning

confidence: 99%

“…As we explain above, the problem of finding the optimal autoencoder is equivalent to the problem of Bayesian persuasion and optimal information design (see Kamenica and Gentzkow (2011), Kamenica (2019), Kamenica et al (2021), Bergemann and Morris (2016) for an overview). Several papers study the problem of Bayesian persuasion in the one-dimensional case (i.e., when L = 1 so that X ∈ R 1 ) and derive conditions under which the optimal signal structure is a monotone partition into intervals.…”

Section: G1 the Structure Of Optimal Partitionsmentioning

confidence: 99%

“… See Bronstein et al (2021) for an overview of the important role played by geometric structures in modern deep learning.7 See,Rayo and Segal (2010),Kamenica and Gentzkow (2011), as well as Bergemann and Morris (2019),Kamenica et al (2021) for excellent recent reviews of the subject.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Benign Autoencoders

Malamud¹,

Schrimpf²,

Xu³

et al. 2022

Preprint

View full text Add to dashboard Cite

The success of modern machine learning algorithms depends crucially on efficient data representation and compression through dimensionality reduction. This practice seemingly contradicts the conventional intuition suggesting that data processing always leads to information loss. We prove that this intuition is wrong. For any non-convex problem, there exists an optimal, benign autoencoder (BAE) extracting a lower-dimensional data representation that is strictly beneficial: Compressing model inputs improves model performance. We prove that BAE projects data onto a manifold whose dimension is the compressibility dimension of the learning problem. We develop and implement an efficient algorithm for computing BAE and show that BAE improves model performance in every dataset we consider. Furthermore, by compressing "malignant" data dimensions, BAE makes learning more stable and robust.

show abstract

Section: E2 Generalized Optimal Autoencoder Formulationmentioning

confidence: 99%

Section: G1 the Structure Of Optimal Partitionsmentioning

confidence: 99%

See 1 more Smart Citation

Benign Autoencoders

Malamud¹,

Schrimpf²,

Xu³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…22 Some of this work is surveyed by Kamenica, Kim, and Zapechelnyuk (2021). Calzolari and Pavan (2006a,b), Pavan and Calzolari (2009), Rosar (2017), Georgiadis and Szentes (2020), Dworczak (2020), Doval and Skreta (2021) and Boleslavsky and Kim (2021) study persuasion problems subject to incentive constraints. …”

mentioning

confidence: 99%

“…Some of this work is surveyed byKamenica, Kim, and Zapechelnyuk (2021).Calzolari and Pavan (2006a,b),Pavan and Calzolari (2009),Rosar (2017),Georgiadis and Szentes (2020),Dworczak (2020),Doval and Skreta (2021) andBoleslavsky and Kim (2021) study persuasion problems subject to incentive constraints.3Brocas and Carrillo (2007) study a similar setting in discrete time, with no discounting. See alsoFershtman and Pavan (2022, §4 and §5).…”

mentioning

confidence: 99%

Slow persuasion

Escudé

Sinander

2023

View full text Add to dashboard Cite

What are the value and form of optimal persuasion when information can be generated only slowly? We study this question in a dynamic model in which a “sender” provides public information over time subject to a graduality constraint, and a decision maker takes an action in each period. Using a novel “viscosity” dynamic programming principle, we characterize the sender's equilibrium value function and information provision. We show that the graduality constraint inhibits information provision relative to unconstrained persuasion. The gap can be substantial, but closes as the constraint slackens. Contrary to unconstrained persuasion, less‐than‐full information may be provided even if players have aligned preferences but different prior beliefs.

show abstract

On the equivalence of information design by uninformed and informed principals

Zapechelnyuk

2023

Econ Theory

Self Cite

View full text Add to dashboard Cite

We compare information design, or Bayesian persuasion, by an uninformed principal (who has no information about the state of the world when making her choice) and by an informed principal (who has private information and can condition her choice on that information). We show that, under the assumptions of monotone preferences of the principal and nondegenerate information structures, a Pareto undominated outcome is implementable by the uninformed principal if and only if it is implementable by the informed principal.

show abstract

Bayesian persuasion and information design: perspectives and open issues

Cited by 7 publications

References 35 publications

Benign Autoencoders

Benign Autoencoders

Slow persuasion

On the equivalence of information design by uninformed and informed principals

Contact Info

Product

Resources

About