arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1901.11271cs.NE
收藏

用生成式神经网络改进进化策略

Improving Evolutionary Strategies with Generative Neural Networks

Louis Faury, Clement Calauzenes, Olivier Fercoq, Syrine Krichen

摘要:

进化策略(ES)是一种流行的黑盒零阶优化算法,它依赖于搜索分布来有效地优化各种目标函数。本文研究了在经典ES算法中使用高度灵活的搜索分布的潜在好处,与标准的(通常是高斯)。我们用生成神经网络(GNNs)对这种分布进行建模,并引入了一种新的训练算法,该算法利用它们的表现力来加速ES过程。我们表明,这种定制的算法可以很容易地将现有的ES算法,并优于国家的最先进的不同的目标函数。

英文摘要:

Evolutionary Strategies (ES) are a popular family of black-box zeroth-order optimization algorithms which rely on search distributions to efficiently optimize a large variety of objective functions. This paper investigates the potential benefits of using highly flexible search distributions in classical ES algorithms, in contrast to standard ones (typically Gaussians). We model such distributions with Generative Neural Networks (GNNs) and introduce a new training algorithm that leverages their expressiveness to accelerate the ES procedure. We show that this tailored algorithm can readily incorporate existing ES algorithms, and outperforms the state-of-the-art on diverse objective functions.

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。

↑