Building computer vision systems using machine learning algorithms

• Autorzy: Boyko N. , Dokhniak B. , Korkishko V.
• Języki publikacji: EN

Abstrakty

EN

This article is devoted to the algorithm of training with reinforcement (reinforcement learning). This article will cover various modifications of the Q-Learning algorithm, along with its techniques, which can accelerate learning using neural networks. We also talk about different ways of approximating the tables of this algorithm, consider its implementation in the code and analyze its behavior in different environments. We set the optimal parameters for its implementation, and we will evaluate its performance in two parameters: the number of necessary neural network weight corrections and quality of training.

Słowa kluczowe

EN

training with reinforcement; Q-Learning; neural networks; Markov environment

• Wydawca: Polish Academy of Sciences, Branch in Lublin
• Czasopismo: ECONTECHMOD : An International Quarterly Journal on Economics of Technology and Modelling Processes
• Rocznik: 2018
• Tom: Vol. 7, No 2
• Strony: 9--14
• Opis fizyczny: Bibliogr. 20 poz., rys., wz.

Twórcy

autor

Boyko N.

• Lviv Polytechnic National University, Lviv, Ukraine

autor

Dokhniak B.

• Lviv Polytechnic National University, Lviv, Ukraine

autor

Korkishko V.

• Lviv Polytechnic National University, Lviv, Ukraine

Bibliografia

• 1. Boyko N. 2016. Basic concepts of dynamic recurrent neural networks development / N. Boyko, P. Pobereyko // ECONTECHMOD : an international quarterly journal on economics of technology and modelling processes. – Lublin: Polish Academy of Sciences. – Vol. 5, № 2. – P. 63-68.
• 2. Coelho L. 2013. Building Machine Learning Systems with Python/ Luis Pedro Coelho, Willi Richert. – Birmingham –Mumbai: Published by Packt Publishing Ltd. – 290 p.
• 3. Bishop C . M. 2006. Pattern recognition and machine learning / Christopher M. Bishop. – Springer Science+Business Media, LLC. – 78 p.
• 4. Elkan C. 2003. Using the triangle inequality to accelebrate k-means / C. Elkan // In Proceedings of the Twelfth International Conference on Machine Learning, 2003. – P. 147–153.
• 5. Matov O. Ia. 2009. Modern technologies of information resources integration / O. Ia. Matov // Registration, storage and processing of data. - V. 11, № 1, P. 33–42.
• 6. Khramova I. O. 2009. The use of service-oriented architectures in the integration of information resources / I. O. Khramova // Registration, storage and processing of data. - V. 11, № 2, P. 70–76.
• 7. Matov O. Ia. 2009. Mathematical models of conflict losses performance of the mediators ontology for General use in GRID environment / O. Ia. Matov // Registration, storage and processing of data. -V. 11, № 3, P. 18–25.
• 8. Matov O. Ia. 2007. The problem of horizontal integration of information resources in a multi-tiered organizational structures with dynamic configuration / O. Ia. Matov // Registration, storage and processing of data. - V. 9, № 3, P. 88–97.
• 9. Matov O. Ia. 2006. Dynamic integration of information resources of the unified information infrastructure of the electricity market / O.Ia. Matov // The functioning and development of electricity and gas markets: collection ofscientific works Institute of modelling in energy im. H.Ie. Pukhova. - P. 93–98.
• 10. Boyko N. 2016. A look trough methods of intellectual data analysis and their applying in informational systems / N. Boyko // Computer sciences and informatopn technologies CSIT 2016 : Proceedings of XI International scientific conference CSIT 2016 : proceedings. – Lviv: Publ ofv Lviv Polytechnik. – P. 183-185.
• 11. Boyko N. 2016. Basic concepts of dynamic recurrent neural networks development / N. Boyko, P. Pobereyko // ECONTECHMOD: an international quarterly journal on economics of technology and modelling processes. – Lublin: Polish Academy of Sciences, Vol. 5, № 2. – P. 63-68.
• 12. Leskovec J. 2014. Mining of massive datasets / J. Leskovec, A. Rajaraman, J. D. Ullman. – Massachusetts: Cambridge University Press. – 470 р.
• 13. Boyko N. 2017. Use of a cloud storage for implementation informational proce / N. Boyko // ECONTECHMOD : an international quarterly journal on economics of technology and modelling processes –Vol. 2 No.6. –Branch in Lublin: Polish Academy of Sciences, 2017. – P. 3-8.
• 14. Boyko N. 2017. Building computer vision systems using machine learning algorithms / N. Boyko, N. Sokil // ECONTECHMOD : an international quarterly journal on economics of technology and modelling processes – Vol. 2 No.6. – Branch in Lublin: Polish Academy of Sciences, 2017. – P. 15-20.
• 15. Boyko N. 2016. Using genetic algorithms for modeling informational processes / N. Boyko // Computational problems of electrical engineering : scientific journal "Computational problems elektotehniky" –Vol. 6 No. 1(10) – Founder and Publisher Lviv Polytechnic National University, 2016. – P. 55-62.
• 16. Boyko N. 2016. Application of mathematical models for improvement of “cloud” data processes organization / N. Boyko // Mathematical Modeling and Computing : scientific journal "Computational problems elektotehniky" – Vol. 3 No. 2 – Founder and Publisher Lviv Polytechnic National University, 2016. – P. 111-119.
• 17. Boyko N. I. 2017. The technological capabilities of the Hyperwave / NI information server. Boyko, O. V. Kopach // International Scientific and Practical Conference "Information Technologies and Computer Modeling", May 15-20, 2017: Abstracts / Repr. for the issue Volodarsky Ye.T. - Ivano-Frankivsk: Mr. Golin O. M. - P. 8-11 p.
• 18. Boyko N. I. 2017. Perspective technologies of research of large data in distributed information systems / N. I. Boyko // Radioelectronics, computer science, management. № 4. - Zaporozhye: Zaporizhzhya National Technical University. - P. 66-77.
• 19. Maass W. 2002. Real-time computing without stable states: a new framework for neural computations based on perturbations / W. Maass, T. Natschger, H. Markram / Neural Computation : proceedings. –Switzerland: Institute for Theoretical Computer Science, Vol. 11. – P. 2531–2560.
• 20. Schrauwen B., Verstraeten D., Campenhout J. V. 2007. An overview of reservoir computing theory, applications and implementations / B. Schrauwen, D. Verstraeten, J. V. Campenhout // Proc. of the 15th European Symp. on Artificial Neural Networks : proceedings. – Belgium: Bruges,. P. 471–482.

Uwagi

Opracowanie rekordu w ramach umowy 509/P-DUN/2018 ze środków MNiSW przeznaczonych na działalność upowszechniającą naukę (2018).