Optimizing Regret
Building on the identity that expected regret equals the covariance between costs and decisions, this paper develops a derivative theory of the covariance regret functional. We derive the Gâteaux derivative, showing that the universal steepest-descent direction is the contrarian policy $-(c-\bar c)$, while ascent yields momentum. For linear policies $\hatπ(c)=Ac+b$, the gradient is the cost covariance matrix $Σ_c$, with a zero Hessian implying boundary-optimal solutions such as the minimum-variance portfolio. We extend to constrained optimization, sign-gradient duality between regret minimization and alpha maximization, finite-sample convergence bounds paralleling Thompson Sampling, and gradient-descent algorithms requiring only input observations.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Optimizing Through Change: Bounds and Recommendations for Time-Varying Bayesian Optimization Algorithms
Time-Varying Bayesian Optimization (TVBO) is the go-to framework for optimizing a time-varying, expensive, noisy black-box function. However, most of the solutions proposed so far either rely on unrealistic assumptions o…
Bayesian OptimizationAsymptotic Performance of Time-Varying Bayesian Optimization
Time-Varying Bayesian Optimization (TVBO) is the go-to framework for optimizing a time-varying black-box objective function that may be noisy and expensive to evaluate. Is it possible for the instantaneous regret of a TV…
Bayesian OptimizationEmpirical Bayes Regret Minimization
Most bandit algorithm designs are purely theoretical. Therefore, they have strong regret guarantees, but also are often too conservative in practice. In this work, we pioneer the idea of algorithm design by minimizing th…
Regret Optimality of GP-UCB
Gaussian Process Upper Confidence Bound (GP-UCB) is one of the most popular methods for optimizing black-box functions with noisy observations, due to its simple structure and superior performance. Its empirical successe…
Bayesian OptimizationSmart Predict-Then-Control: Integrating identification and control via decision regret
This paper presents Smart Predict-Then-Control (SPC) framework for integrating system identification and control. This novel SPC framework addresses the limitations of traditional methods, the unaligned modeling error an…