arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1112.4454cs.NEcs.NAmath.NAquant-ph
收藏

进化Hessian学习:强制最优协方差自适应学习(FOCAL)

Evolutionary Hessian Learning: Forced Optimal Covariance Adaptive Learning (FOCAL)

Princeton University(普林斯顿大学);Colorado State University(科罗拉多州立大学)

Ofer M. Shir, Jonathan Roslund, Darrell Whitley, Herschel Rabitz

AI总结:

针对CMA-ES在条件数大的高维景观中无法学习真实协方差矩阵的问题,提出FOCAL方法,通过强制学习全局吸引盆处的Hessian矩阵,在模型景观和量子控制系统中高保真恢复Hessian。

摘要:

协方差矩阵自适应进化策略(CMA-ES)一直是利用协方差信息最成功的进化策略;它使用一种主成分分析形式,在某些条件下,被建议收敛到正确的协方差矩阵,该矩阵被表述为数学上定义良好的Hessian矩阵的逆。然而,在实践中,存在CMA-ES收敛到全局最优(完成其主要目标)但未学习真实协方差矩阵(错过辅助目标)的情况,这可能是由于步长不足。这些情况可能涉及具有大条件数的高维景观。本文介绍了一种名为强制最优协方差自适应学习(FOCAL)的新技术,其明确目标是确定全局吸引盆处的Hessian。首先介绍了学习协方差与Hessian矩阵之间逆关系的理论基础。然后引入FOCAL,并证明其在模型景观和实验量子控制系统(观察到具有不可分离、非二次搜索景观)上能够高保真地恢复Hessian矩阵。恢复的Hessian形式通过系统的物理知识得到证实。这项研究构成了自然计算成功服务于其他自然科学分支的一个例子,同时为任何寻求景观信息的高维连续搜索引入了一种强大的通用方法。

英文摘要:

The Covariance Matrix Adaptation Evolution Strategy (CMA-ES) has been the most successful Evolution Strategy at exploiting covariance information; it uses a form of Principle Component Analysis which, under certain conditions, is suggested to converge to the correct covariance matrix, formulated as the inverse of the mathematically well-defined Hessian matrix. However, in practice, there exist conditions where CMA-ES converges to the global optimum (accomplishing its primary goal) while it does not learn the true covariance matrix (missing an auxiliary objective), likely due to step-size deficiency. These circumstances can involve high-dimensional landscapes with large condition numbers. This paper introduces a novel technique entitled Forced Optimal Covariance Adaptive Learning (FOCAL), with the explicit goal of determining the Hessian at the global basin of attraction. It begins by introducing theoretical foundations to the inverse relationship between the learned covariance and the Hessian matrices. FOCAL is then introduced and demonstrated to retrieve the Hessian matrix with high fidelity on both model landscapes and experimental Quantum Control systems, which are observed to possess a non-separable, non-quadratic search landscape. The recovered Hessian forms are corroborated by physical knowledge of the systems. This study constitutes an example for Natural Computing successfully serving other branches of natural sciences, and introducing at the same time a powerful generic method for any high-dimensional continuous search seeking landscape information.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑