一种为复杂系统提供深入了解并为不透明的机器学习算法提供性能基准的透明开放箱式学习网络

原文网址：http://www.astp-agr.com/index.php/Index/Index/detail?id=57

         如今，普遍通过部署神经网络和机器学习算法来提供从具有多个基本变量的复杂系统中得出的预测是司空见惯的。这在特别针对在能直接测量的关键变量的直接测量数量有限或难以获得的情况下下特别有用。有许多石油系统属于这种情况适合这种描述的石油系统。尽管人工智能算法提供了有效的解决方案来预测这些难以衡量的因变量，但它们往往无法提供对底层系统的洞察力更深入的了解，且而用于推导其预测的变量之间的关系并不确定。对用户来说，这样的系统通常代表“黑匣子”。

       最近，英国DWA能源的David A.Wood教授研究发现新型透明开放盒箱式（TOB）学习网络算法通过清楚地揭示其中间计算和应用于其独立变量可以得出其预测的权重来克服这些限制。本文详细描述了构建和优化TOB网络所涉及的步骤。对于中小型数据集，可以使用电子表格公式和标准优化器部署TOB网络；对于较大的系统，编码算法和定制优化器易于配置。将电子表格优点（如Microsoft Excel解算器）与算法代码相结合的混合应用程序也是有效的。 TOB学习网络适用于三大油田数据集，并展示了其学习能力以及对其能够提供的建模系统的洞察力了解。但 TOB不是用来替代神经网络和机器学习算法，而是作为补充工具为一些不太透明的算法提供的性能基准。

A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms

David A. Wood

(Published: 2018-03-20)

CorrespondingAuthor and Email: David A. Wood, dw@dwasolutions.com

Citation: Wood, D.A. A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms. Advances in Geo-Energy Research, 2018, 2(2): 148-162, doi: 10.26804/ager.2018.02.04.

ArticleType: Original article

Abstract：

    It is now commonplace to deploy neural networks and machine-learning algorithms to provide predictions derived from complex systems with multiple underlying variables. This is particularly useful where direct measurements for the key variables are limited in number and/or  difficult  to  obtain.  There  are  many  petroleum  systems  that  fit  this  description. Whereas  artificial  intelligence  algorithms  offer  effective  solutions  to  predicting  these difficult-to-measure dependent variables they often fail to provide insight to the underlying systems and the relationships between the variables used to derive their predictions areobscure. To the user such systems often represent “black boxe”.

    The novel transparent open box (TOB) learning network algorithm described here overcomes these limitations by clearly revealing its intermediate calculations and the weightings applied to its independent variables in deriving its predictions. The steps involved in building and optimizing the TOB network are described in detail. For small to mid-sized datasets the TOB network can be deployed using spreadsheet formulas and standard optimizers; for larger systems coded algorithms and customised optimizers are easy to configure. Hybrid applications combining spreadsheet benefits (e.g. Microsoft Excel Solver) with algorithm code are also effective. The TOB learning network is applied to three petroleum datasets and demonstrates both its learning capabilities and the insight to the modelled systems that it is able to provide. TOB is not proposed as a replacement for neural networks and machine learning algorithms, but as a complementary tool; it can serve as a performance benchmark for some of the less transparent algorithms.

Keywords: Learning networks, transparency of variable relationships, benchmarking machine learning, performance, prediction of complex petroleum systems, soft-computing solutions, under-fitting/over-fitting.

PDF下载：5b124d2e22c56.pdf