Using Modern Architectures of Recurrent Neural Networks for Technical Diagnosis of Complex Systems

The paper provides an overview of modern recurrent neural network architectures that can be used as a model (building tool) for diagnosis and predicting of infocommunication systems elements. The paper presents arguments in favor of the using of recurrent neural networks. Also, to determine the best architecture, were conducted a number of experiments, and on the basis of the results, conclusions were made about the future prospects of using modern architecture.

[1]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[3]  P J Webros.  BACKPROPAGATION THROUGH TIME: WHAT IT DOES AND HOW TO DO IT , 1990 .

[4]  Александр Викторович Близняков,et al.  Increase effectiveness of reversible braking mode realization of the wound-rotor induction motor , 2015 .

[5]  E. B. Kolesnikov,et al.  System of neural network control and prediction technical conditions electromechanical systems , 2016, 2016 2nd International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM).

[6]  Andrii Oliinyk,et al.  Development of the method for decomposition of superpositions of unknown pulsed signals using the second­order adaptive spectral analysis , 2018 .

[7]  Yoshua Bengio,et al.  Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling , 2014, ArXiv.

[8]  Sepp Hochreiter,et al.  Untersuchungen zu dynamischen neuronalen Netzen , 1991 .

[9]  Andrii Oliinyk,et al.  Synthesis of Neuro-Fuzzy Networks on the Basis of Association Rules , 2014 .

[10]  M. Kotsur,et al.  An application of scheme and field models for simulation of electromagnetic processes of power transformers , 2018, 2018 14th International Conference on Advanced Trends in Radioelecrtronics, Telecommunications and Computer Engineering (TCSET).

[11]  Yoshua Bengio,et al.  Diffusion of Credit in Markovian Models , 1994, NIPS.

[12]  Marek Kowal,et al.  Neuro-fuzzy networks and their application to fault detection of dynamical systems , 2007, Eng. Appl. Artif. Intell..

[13]  Yoshua Bengio,et al.  Learning long-term dependencies with gradient descent is difficult , 1994, IEEE Trans. Neural Networks.

[14]  Charles Elkan,et al.  Learning to Diagnose with LSTM Recurrent Neural Networks , 2015, ICLR.

[15]  V. Lovkin,et al.  Improved method of group decision making in expert systems based on competitive agents selection , 2017, 2017 IEEE First Ukraine Conference on Electrical and Computer Engineering (UKRCON).

[16]  Katsuba Yurii,et al.  Application of Artificial Neural Networks in Vehicles’ Design Self-Diagnostic Systems for Safety Reasons , 2017 .

[17]  B. B. Samotokin,et al.  DEVELOPMENT OF STRATIFIED APPROACH TO SOFTWARE DEFINED NETWORKS SIMULATION , 2017 .