Definition
Bayesian regret quantifies the difference between the optimal reward achievable with perfect information and the expected reward obtained by an agent acting under uncertainty. It is calculated by integrating the regret over all possible states of the world weighted by their prior probabilities. This concept is crucial in reinforcement learning and game theory, helping to evaluate how well an algorithm performs when it must make decisions without knowing the true underlying parameters or environment dynamics.
Summary
A metric in decision theory measuring the expected loss due to uncertainty about the true state of the world.
Key Concepts
- Expected utility
- Prior distribution
- Decision theory
- Regret minimization
Use Cases
- Reinforcement learning evaluation
- Multi-armed bandit problems
- Strategic game analysis