The Hazards and Benefits of Condescension in Social Learning

01/26/2023

∙

In a misspecified social learning setting, agents are condescending if they perceive their peers as having private information that is of lower quality than it is in reality. Applying this to a standard sequential model, we show that outcomes improve when agents are mildly condescending. In contrast, too much condescension leads to bad outcomes, as does anti-condescension.

READ FULL TEXT

The Hazards and Benefits of Condescension in Social Learning

Inducing Cooperative behaviour in Sequential-Social dilemmas through Multi-Agent Reinforcement Learning using Status-Quo Loss

Improving Social Welfare While Preserving Autonomy via a Pareto Mediator

Bayesian Social Learning in a Dynamic Environment

More Specificity, More Attention to Social Context: Reframing How We Address "Bad Actors"

Computers as Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially

Overabundant Information and Learning Traps

The Iterated Local Model for Social Networks

The Hazards and Benefits of Condescension in Social Learning

Related Research

Inducing Cooperative behaviour in Sequential-Social dilemmas through Multi-Agent Reinforcement Learning using Status-Quo Loss

Improving Social Welfare While Preserving Autonomy via a Pareto Mediator

Bayesian Social Learning in a Dynamic Environment

More Specificity, More Attention to Social Context: Reframing How We Address "Bad Actors"

Computers as Bad Social Actors: Dark Patterns and Anti-Patterns in Interfaces that Act Socially

Overabundant Information and Learning Traps

The Iterated Local Model for Social Networks