Generalised Entropy MDPs and Minimax Regret

12/10/2014

∙

Bayesian methods suffer from the problem of how to specify prior beliefs. One interesting idea is to consider worst-case priors. This requires solving a stochastic zero-sum game. In this paper, we extend well-known results from bandit theory in order to discover minimax-Bayes policies and discuss when they are practical.

READ FULL TEXT

Generalised Entropy MDPs and Minimax Regret

On Worst-case Regret of Linear Thompson Sampling

An Information-Theoretic Approach to Minimax Regret in Partial Monitoring

Expected Worst Case Regret via Stochastic Sequential Covering

Leveraging vague prior information in general models via iteratively constructed Gamma-minimax estimators

Learning Minimax Estimators via Online Learning

Teaching decision theory proof strategies using a crowdsourcing problem

Gaussian One-Armed Bandit and Optimization of Batch Data Processing

Generalised Entropy MDPs and Minimax Regret

Related Research

On Worst-case Regret of Linear Thompson Sampling

An Information-Theoretic Approach to Minimax Regret in Partial Monitoring

Expected Worst Case Regret via Stochastic Sequential Covering

Leveraging vague prior information in general models via iteratively constructed Gamma-minimax estimators

Learning Minimax Estimators via Online Learning

Teaching decision theory proof strategies using a crowdsourcing problem

Gaussian One-Armed Bandit and Optimization of Batch Data Processing