arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1408.4622stat.COcs.LGmath.OCstat.ML
收藏

一种新的贝叶斯优化积分损失函数

A new integral loss function for Bayesian optimization

SUPELEC(高等电力学院)

Emmanuel Vazquez, Julien Bect

摘要:

我们考虑的问题,最大化的实值连续函数$f$使用贝叶斯方法。自70年代Jonas Mockus和Antanas Mockilinskas的早期工作以来,优化问题通常通过考虑损失函数$\max f - M_n$(其中$M_n$表示在对$f$进行$n$评估后观察到的最佳函数值)来公式化。这个损失函数强调最大值的价值,而牺牲了最大化者的位置。在一步贝叶斯最优策略的特殊情况下,它导致经典的期望改进(EI)抽样标准。这是逐步不确定性降低(SUR)策略的特殊情况,其中与感兴趣的数量上的特定不确定性度量(此处为预期损失)相关的风险在算法的每一步都被最小化。在这篇文章中,假设$f$是定义在测量空间$(\mathbb{X}, λ)$,我们建议,而不是考虑的积分损失函数$\int_{\mathbb{X}} (f - M_n)_{+}\, dλ$,我们表明,这导致,在高斯过程之前的情况下,到一个新的数值上易于处理的采样标准,我们称之为$\rm EI^2$(预期的综合预期改善)。数值实验表明,基于这种新的采样准则的SUR策略减少了错误的值和位置的最大化比基于EI的策略。

英文摘要:

We consider the problem of maximizing a real-valued continuous function $f$ using a Bayesian approach. Since the early work of Jonas Mockus and Antanas Žilinskas in the 70's, the problem of optimization is usually formulated by considering the loss function $\max f - M_n$ (where $M_n$ denotes the best function value observed after $n$ evaluations of $f$). This loss function puts emphasis on the value of the maximum, at the expense of the location of the maximizer. In the special case of a one-step Bayes-optimal strategy, it leads to the classical Expected Improvement (EI) sampling criterion. This is a special case of a Stepwise Uncertainty Reduction (SUR) strategy, where the risk associated to a certain uncertainty measure (here, the expected loss) on the quantity of interest is minimized at each step of the algorithm. In this article, assuming that $f$ is defined over a measure space $(\mathbb{X}, λ)$, we propose to consider instead the integral loss function $\int_{\mathbb{X}} (f - M_n)_{+}\, dλ$, and we show that this leads, in the case of a Gaussian process prior, to a new numerically tractable sampling criterion that we call $\rm EI^2$ (for Expected Integrated Expected Improvement). A numerical experiment illustrates that a SUR strategy based on this new sampling criterion reduces the error on both the value and the location of the maximizer faster than the EI-based strategy.

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。

↑