Open Access
Issue
Published:
20 March 2018
A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms
(
)
Download citation
970
Views
60
Downloads
48
Crossref
0
WoS
51
Scopus
N/A
CSCD
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 are obscure. 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.
menu
A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms
(
)
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 are obscure. 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.
References(28)
Aalst, W.M.P., Rubin, V., Verbeek, H.M.W., et al. Process mining: A two-step approach to balance between underfitting and overfitting. Softw. Syst. Model. 2010, 9(1): 87-111.
Behnke, S. Hierarchical neural networks for image interpretation. Lect. Notes Comput. Sci. 2003, 2766(3): 1345-1346
Choubineh, A., Ghorbani, H., Wood, D.A., et al. Improved predictions of wellhead choke liquid critical-flow rates: modelling based on hybrid neural network training learning based optimization. Fuel 2017, 207: 547-560.
Cortes, C., Vapnik, V.N. Support-vector networks. Mach. Learn 1995, 20(3): 273-297.
Cybenko, G. Approximation by superpositions of a sigmoidal function. Math. Control Signal. 1989, 2(4): 303-314.
Elkatatny, S., Tariq, Z., Mahmoud, M. Real time prediction of drilling fluid rheological properties using Artificial Neural Networks visible mathematical model (white box). J. Pet. Sci. Eng. 2016, 146: 1202-1210.
Farley, B., Clark, W.A. Simulation of self-organizing systems by digital computer. IEEE Trans. Inf. Theory 1954, 4(4): 76-84.
Gomaa, S. New bubble point pressure correlation for middle east crude oils. Int. Adv. Res. J. Sci. Eng. Technol. 2016, 3(12): 1-9.
Heinert, M. Artificial neural networks-how to open the black boxes. Application of Artificial Intelligence in Engineering Geodesy, 2008.
Ince, D.C. Collected Works of AM Turing-Mechanical Intelligence. Mechanical intelligence, 1992.
Jang, J.S. ANFIS: Adaptive-network-based fuzzy inference system. IEEE Trans. Syst. Man. Cybern. Syst. 1993, 23(3): 665-685.
Li, Y., Xie, R., Yu, L. Flow pattern identification of oil-gas-water three-phase flow based on NPSO-LSSVM algorithm. J. Theor. Appl. Inf. Technol. 2013, 48(2): 933-938.
McCulloch, W., Pitts, W. A logical calculus of ideas immanent in nervous activity. Bull. Math. Biol. 1943, 5(4): 115-133.
Meng, M., Zhao, C. Application of support vector machines to a small-sample prediction. Adv. Pet. Explor. Dev. 2015, 10(2): 72-75.
Rosenblatt, F. The Perceptron: A probabilistic model for information storage and organization in the brain. Psychol. Rev. 1958, 65(6): 386-408.
Santos, R.B., Rupp, M., Bonzi, S.J., et al. Comparison between multilayer feedforward neural networks and a radial basis function network to detect and locate leaks in pipelines transporting gas. Chem. Eng. Trans. 2013, 32: 1375-1380.
Scarselli, F., Tsoil, A.C. Universal approximation using feedforward neural networks: A survey of some existing methods, and some new results. Neural Netw. 1998, 11(1): 15-37.
Schmidhuber, J. Learning complex, extended sequences using the principle of history compression. Neural Comput. 1992, 4(2): 234-242.
Schmidhuber, J. Deep learning in neural networks: An overview. Neural Netw. 2015, 61: 85-117.
Sheremetov, L., Batyrsgin, I., Filatov, D., et al. Fuzzy expert sSystem for solving lost circulation problem. Appl. Soft Comput. 2008, 8(1): 14-29.
Sugeno, M., Kang, G. Structure identification of fuzzy model. Fuzzy Sets. Syst. 1988, 28(1): 15-33.
Suykens, J.A.K., Vandewalle, J. Least squares support vector machine classifiers. Neural Process. Lett. 1999, 9(3): 293-300.
Werbos, P.J. Beyond regression: New tools for prediction and analysis in the behavioral sciences. Boston, Harvard University, 1975.
Yavari, H., Sabah, M., Khosravanian, R., et al. Application of adaptive neuro-fuzzy inference system and mathematical ROP models for prediction of drilling rate. J. Oil Gas Sci. Technol. 2018.
Zamani, H.A., Rafiee Taghanaki, S., Karimi, M., et al. Implementing ANFIS for prediction of reservoir oil solution gas-oil ratio. J. Nat. Gas Sci. Eng. 2015, 25: 325-334.
Publication history
Copyright
Acknowledgements
This work is influenced by my interactions with many researchers around the world on machine learning algorithms over many years. That has stimulated my desire to develop more transparent learning-based algorithms.
Rights and permissions
This article is distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
FEEDBACK 
TOP