{"id":1145,"date":"2021-01-15T21:57:43","date_gmt":"2021-01-15T19:57:43","guid":{"rendered":"http:\/\/journals.khnu.km.ua\/vestnik\/?p=1145"},"modified":"2021-03-23T11:35:02","modified_gmt":"2021-03-23T09:35:02","slug":"application-of-a-genetic-algorithm-to-search-for-the-optimal-convolutional-neural-network-architecture-with-weight-distribution","status":"publish","type":"post","link":"https:\/\/journals.khnu.km.ua\/vestnik\/?p=1145","title":{"rendered":"Application of a genetic algorithm to search for the optimal convolutional neural network architecture with weight distribution"},"content":{"rendered":"

APPLICATION OF A GENETIC ALGORITHM TO SEARCH FOR THE OPTIMAL CONVOLUTIONAL NEURAL NETWORK ARCHITECTURE WITH WEIGHT DISTRIBUTION<\/p>\n

\u0417\u0410\u0421\u0422\u041e\u0421\u0423\u0412\u0410\u041d\u041d\u042f \u0413\u0415\u041d\u0415\u0422\u0418\u0427\u041d\u041e\u0413\u041e \u0410\u041b\u0413\u041e\u0420\u0418\u0422\u041c\u0423 \u0414\u041b\u042f \u041f\u041e\u0428\u0423\u041a\u0423 \u041e\u041f\u0422\u0418\u041c\u0410\u041b\u042c\u041d\u041e\u0407 \u0410\u0420\u0425\u0406\u0422\u0415\u041a\u0422\u0423\u0420\u0418 \u0417\u0413\u041e\u0420\u0422\u041a\u041e\u0412\u041e\u0407 \u041d\u0415\u0419\u0420\u041e\u041d\u041d\u041e\u0407 \u041c\u0415\u0420\u0415\u0416\u0406 \u0417 \u0420\u041e\u0417\u041f\u041e\u0414\u0406\u041b\u0415\u041d\u041d\u042f\u041c \u0412\u0410\u0413<\/p>\n

<\/a> \u0421\u0442\u043e\u0440\u0456\u043d\u043a\u0438: 7-11. \u041d\u043e\u043c\u0435\u0440: \u21161, 2020 (281)<\/strong>
\n\u0410\u0432\u0442\u043e\u0440\u0438:<\/strong>
\nP.M. RADIUK
\nKhmelnytskyi National University
\n\u041f.\u041c. \u0420\u0410\u0414\u042e\u041a
\n\u0425\u043c\u0435\u043b\u044c\u043d\u0438\u0446\u044c\u043a\u0438\u0439 \u043d\u0430\u0446\u0456\u043e\u043d\u0430\u043b\u044c\u043d\u0438\u0439 \u0443\u043d\u0456\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442
\nDOI:<\/strong> https:\/\/www.doi.org\/10.31891\/2307-5732-2020-281-1-7-11<\/a>
\n\u0420\u0435\u0446\u0435\u043d\u0437\u0456\u044f\/Peer review :<\/strong> 04. 01.2020 \u0440.
\n\u041d\u0430\u0434\u0440\u0443\u043a\u043e\u0432\u0430\u043d\u0430\/Printed :<\/strong> 14.02.2020 \u0440.<\/p>\n

\u0410\u043d\u043e\u0442\u0430\u0446\u0456\u044f \u043c\u043e\u0432\u043e\u044e \u043e\u0440\u0438\u0433\u0456\u043d\u0430\u043b\u0443<\/strong><\/p>\n

In the past decade, a new way in neural networks research called Network architectures search has demonstrated noticeable results in the design of architectures for image segmentation and classification. Despite the considerable success of the architecture search in image segmentation and classification, it is still an unresolved and urgent problem. Moreover, the neural architecture search is also a highly computationally expensive task. This work proposes a new approach based on a genetic algorithm to search for the optimal convolutional neural network architecture. We integrated a genetic algorithm with standard stochastic gradient descent that implements weight distribution across all architecture solutions. This approach utilises a genetic algorithm to design a sub-graph of a convolution cell, which maximises the accuracy on the validation set. We show the performance of our approach on the CIFAR-10 and CIFAR-100 datasets with a final accuracy of 93.21% and 78.89%, respectively. The main scientific contribution of our work is the combination of genetic algorithm with weight distribution in the architecture search tasks that achieve similar to state-of-the-art results on a single GPU.
\nKeywords:<\/strong> convolutional neural networks, genetic algorithms, weight distribution, ablation study.<\/p>\n

\u0420\u043e\u0437\u0448\u0438\u0440\u0435\u043d\u0430 \u0430\u043d\u043e\u0442\u0430\u0446\u0456\u044f \u0430\u043d\u0433\u043b\u0456\u0439\u0441\u044c\u043a\u043e\u044e \u043c\u043e\u0432\u043e\u044e<\/strong><\/p>\n

\u0417\u0430 \u043e\u0441\u0442\u0430\u043d\u043d\u0454 \u0434\u0435\u0441\u044f\u0442\u0438\u043b\u0456\u0442\u0442\u044f \u043d\u043e\u0432\u0438\u0439 \u0441\u043f\u043e\u0441\u0456\u0431 \u0434\u043e\u0441\u043b\u0456\u0434\u0436\u0435\u043d\u043d\u044f \u043d\u0435\u0439\u0440\u043e\u043d\u043d\u0438\u0445 \u043c\u0435\u0440\u0435\u0436 \u043f\u0456\u0434 \u043d\u0430\u0437\u0432\u043e\u044e \u00ab\u041f\u043e\u0448\u0443\u043a \u043c\u0435\u0440\u0435\u0436\u0435\u0432\u0438\u0445 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440\u00bb \u043f\u0440\u043e\u0434\u0435\u043c\u043e\u043d\u0441\u0442\u0440\u0443\u0432\u0430\u0432 \u043f\u043e\u0437\u0438\u0442\u0438\u0432\u043d\u0456 \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442\u0438 \u0432 \u0440\u043e\u0437\u0440\u043e\u0431\u0446\u0456 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440 \u0434\u043b\u044f \u0441\u0435\u0433\u043c\u0435\u043d\u0442\u0430\u0446\u0456\u0457 \u0442\u0430 \u043a\u043b\u0430\u0441\u0438\u0444\u0456\u043a\u0430\u0446\u0456\u0457 \u0437\u043e\u0431\u0440\u0430\u0436\u0435\u043d\u044c. \u041d\u0435\u0437\u0432\u0430\u0436\u0430\u044e\u0447\u0438 \u043d\u0430 \u0437\u043d\u0430\u0447\u043d\u0438\u0439 \u0443\u0441\u043f\u0456\u0445 \u043f\u043e\u0448\u0443\u043a\u0443 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440 \u0432 \u0437\u0430\u0434\u0430\u0447\u0430\u0445 \u0441\u0435\u0433\u043c\u0435\u043d\u0442\u0430\u0446\u0456\u0457 \u0442\u0430 \u043a\u043b\u0430\u0441\u0438\u0444\u0456\u043a\u0430\u0446\u0456\u0457 \u0437\u043e\u0431\u0440\u0430\u0436\u0435\u043d\u044c, \u0432\u0456\u043d \u0432\u0441\u0435 \u0449\u0435 \u0454 \u043d\u0435\u0432\u0438\u0440\u0456\u0448\u0435\u043d\u043e\u044e \u0456 \u0430\u043a\u0442\u0443\u0430\u043b\u044c\u043d\u043e\u044e \u043f\u0440\u043e\u0431\u043b\u0435\u043c\u043e\u044e. \u0411\u0456\u043b\u044c\u0448\u0435 \u0442\u043e\u0433\u043e, \u043f\u043e\u0448\u0443\u043a \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440 \u043d\u0435\u0439\u0440\u043e\u043d\u043d\u0438\u0445 \u043c\u0435\u0440\u0435\u0436 \u0454 \u0442\u0430\u043a\u043e\u0436 \u0434\u0443\u0436\u0435 \u0432\u0438\u0442\u0440\u0430\u0442\u0438\u043c \u0437 \u0442\u043e\u0447\u043a\u0438 \u0437\u043e\u0440\u0443 \u043e\u0431\u0447\u0438\u0441\u043b\u044e\u0432\u0430\u043b\u044c\u043d\u0438\u0445 \u0440\u0435\u0441\u0443\u0440\u0441\u0456\u0432. \u0423 \u0446\u0456\u0439 \u0440\u043e\u0431\u043e\u0442\u0456 \u043f\u0440\u043e\u043f\u043e\u043d\u0443\u0454\u0442\u044c\u0441\u044f \u043d\u043e\u0432\u0438\u0439 \u043f\u0456\u0434\u0445\u0456\u0434 \u043d\u0430 \u043e\u0441\u043d\u043e\u0432\u0456 \u0433\u0435\u043d\u0435\u0442\u0438\u0447\u043d\u043e\u0433\u043e \u0430\u043b\u0433\u043e\u0440\u0438\u0442\u043c\u0443 \u0434\u043b\u044f \u043f\u043e\u0448\u0443\u043a\u0443 \u043e\u043f\u0442\u0438\u043c\u0430\u043b\u044c\u043d\u043e\u0457 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440\u0438 \u0437\u0433\u043e\u0440\u0442\u043a\u043e\u0432\u043e\u0457 \u043d\u0435\u0439\u0440\u043e\u043d\u043d\u043e\u0457 \u043c\u0435\u0440\u0435\u0436\u0456. \u041c\u0438 \u0456\u043d\u0442\u0435\u0433\u0440\u0443\u0432\u0430\u043b\u0438 \u0433\u0435\u043d\u0435\u0442\u0438\u0447\u043d\u0438\u0439 \u0430\u043b\u0433\u043e\u0440\u0438\u0442\u043c \u0437\u0456 \u0441\u0442\u0430\u043d\u0434\u0430\u0440\u0442\u043d\u0438\u043c \u0441\u0442\u043e\u0445\u0430\u0441\u0442\u0438\u0447\u043d\u0438\u043c \u0433\u0440\u0430\u0434\u0456\u0454\u043d\u0442\u043e\u043c, \u0449\u043e \u0440\u0435\u0430\u043b\u0456\u0437\u0443\u0454 \u0440\u043e\u0437\u043f\u043e\u0434\u0456\u043b \u0432\u0430\u0433 \u0443 \u0432\u0441\u0456\u0445 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440\u043d\u0438\u0445 \u0440\u0456\u0448\u0435\u043d\u043d\u044f\u0445. \u0426\u0435\u0439 \u043f\u0456\u0434\u0445\u0456\u0434 \u0432\u0438\u043a\u043e\u0440\u0438\u0441\u0442\u043e\u0432\u0443\u0454 \u0433\u0435\u043d\u0435\u0442\u0438\u0447\u043d\u0438\u0439 \u0430\u043b\u0433\u043e\u0440\u0438\u0442\u043c \u0434\u043b\u044f \u043f\u0440\u043e\u0435\u043a\u0442\u0443\u0432\u0430\u043d\u043d\u044f \u0447\u0430\u0441\u0442\u0438\u043d\u0438 \u0433\u0440\u0430\u0444\u0443 \u0432 \u044f\u043a\u043e\u0441\u0442\u0456 \u0437\u0433\u043e\u0440\u0442\u043a\u043e\u0432\u043e\u0433\u043e \u0448\u0430\u0440\u0443, \u0449\u043e \u0437\u0430\u0431\u0435\u0437\u043f\u0435\u0447\u0443\u0454 \u043c\u0430\u043a\u0441\u0438\u043c\u0430\u043b\u044c\u043d\u0443 \u0442\u043e\u0447\u043d\u0456\u0441\u0442\u044c \u043d\u0430 \u0432\u0430\u043b\u0456\u0434\u0430\u0446\u0456\u0439\u043d\u043e\u043c\u0443 \u043d\u0430\u0431\u043e\u0440\u0456 \u0434\u0430\u043d\u0438\u0445. \u0423 \u0446\u0456\u0439 \u0440\u043e\u0431\u043e\u0442\u0456 \u043c\u0438 \u0434\u0435\u043c\u043e\u043d\u0441\u0442\u0440\u0443\u0454\u043c\u043e \u0435\u0444\u0435\u043a\u0442\u0438\u0432\u043d\u0456\u0441\u0442\u044c \u043d\u0430\u0448\u043e\u0433\u043e \u043f\u0456\u0434\u0445\u043e\u0434\u0443 \u043d\u0430 \u043d\u0430\u0431\u043e\u0440\u0430\u0445 \u0434\u0430\u043d\u0438\u0445 CIFAR-10 \u0442\u0430 CIFAR-100 \u0437 \u043a\u0456\u043d\u0446\u0435\u0432\u043e\u044e \u0442\u043e\u0447\u043d\u0456\u0441\u0442\u044e 93,21 % \u0442\u0430 78,89 % \u0432\u0456\u0434\u043f\u043e\u0432\u0456\u0434\u043d\u043e. \u041e\u0441\u043d\u043e\u0432\u043d\u0438\u043c \u043d\u0430\u0443\u043a\u043e\u0432\u0438\u043c \u0432\u043d\u0435\u0441\u043a\u043e\u043c \u043d\u0430\u0448\u043e\u0457 \u0440\u043e\u0431\u043e\u0442\u0438 \u0454 \u043f\u043e\u0454\u0434\u043d\u0430\u043d\u043d\u044f \u0433\u0435\u043d\u0435\u0442\u0438\u0447\u043d\u043e\u0433\u043e \u0430\u043b\u0433\u043e\u0440\u0438\u0442\u043c\u0443 \u0437 \u0440\u043e\u0437\u043f\u043e\u0434\u0456\u043b\u043e\u043c \u0432\u0430\u0433 \u0432 \u0437\u0430\u0434\u0430\u0447\u0430\u0445 \u043f\u043e\u0448\u0443\u043a\u0443 \u0430\u0440\u0445\u0456\u0442\u0435\u043a\u0442\u0443\u0440\u0438, \u0449\u043e \u0434\u043e\u0441\u044f\u0433\u0430\u0454 \u0442\u043e\u0447\u043d\u043e\u0441\u0442\u0456 \u043a\u043b\u0430\u0441\u0438\u0444\u0456\u043a\u0430\u0446\u0457\u0456 \u0437\u043e\u0431\u0440\u0430\u0436\u0435\u043d\u043d\u044f \u0437 \u0432\u0438\u043a\u043e\u0440\u0438\u0441\u0442\u0430\u043d\u043d\u044f\u043c \u043e\u0434\u043d\u043e\u0433\u043e \u0433\u0440\u0430\u0444\u0456\u0447\u043d\u043e\u0433\u043e \u043f\u0440\u043e\u0446\u0435\u0441\u043e\u0440\u0430 \u0431\u043b\u0438\u0437\u044c\u043a\u043e\u0457 \u0434\u043e \u043d\u0430\u0439\u0441\u0443\u0447\u0430\u0441\u043d\u0456\u0448\u0438\u0445 \u0440\u0435\u0437\u0443\u043b\u044c\u0442\u0430\u0442\u0456\u0432.
\n\u041a\u043b\u044e\u0447\u043e\u0432\u0456 \u0441\u043b\u043e\u0432\u0430:<\/strong> \u0437\u0433\u043e\u0440\u0442\u043a\u043e\u0432\u0456 \u043d\u0435\u0439\u0440\u043e\u043d\u043d\u0456 \u043c\u0435\u0440\u0435\u0436\u0456, \u0433\u0435\u043d\u0435\u0442\u0438\u0447\u043d\u0456 \u0430\u043b\u0433\u043e\u0440\u0438\u0442\u043c\u0438, \u0440\u043e\u0437\u043f\u043e\u0434\u0456\u043b\u0435\u043d\u043d\u044f \u0432\u0430\u0433, \u0430\u0431\u043b\u044f\u0446\u0456\u044f \u0434\u043e\u0441\u043b\u0456\u0434\u0436\u0435\u043d\u043d\u044f.<\/p>\n

References<\/strong><\/p>\n

    \n
  1. Romanuke V.V. An efficient technique for size reduction of convolutional neural networks after transfer learning for scene recognition tasks \/ V.V. Romanuke \/\/ Applied Computer Systems. \u2013 2018. \u2013 Volume 23. \u2013 Issue 2. \u2013 P. 141\u2013149. \u2013 DOI: https:\/\/doi.org\/10.2478\/acss-2018-0018<\/li>\n
  2. Zhou H. BayesNAS: A bayesian approach for neural architecture search \/ H. Zhou, M. Yang, J. Wang \/\/ 2019 International Conference on Machine Learning, arXiv:1905.04919 [cs.LG]. \u2013 2019. \u2013 P. 7603\u20137613.<\/li>\n
  3. Savarese P. Learning implicitly recurrent CNNs through parameter sharing \/ P. Savarese, M. Maire \/\/ 2019 International Conference on Learning Representations, arXiv:1902.09701 [cs.LG]. \u2013 2019.<\/li>\n
  4. Pham H. Efficient neural architecture search via parameter sharing \/ H. Pham, M.Y. Guan, B. Zoph \/\/ 35th International Conference on Machine Learning, arXiv:1802.03268 [cs.LG]. \u2013 ICML, 2018. \u2013 Volume 80. \u2013 P. 6522\u20136531.<\/li>\n
  5. Liang H. DARTS+: Improved differentiable architecture search with early stopping \/ H. Liang, S. Zhang, J.\u00a0Sun, X. He, W. Huang, K. Zhuang, Z. Li \/\/ arXiv:1909.06035 [cs.CV]. \u2013 2019.<\/li>\n
  6. Deb K. Introduction to genetic algorithms \/ K. Deb \/\/ Sadhana. \u2013 Academy Proceedings in Engineering Sciences, 1999. \u2013 Volume 24. \u2013 No. 4. \u2013 P. 293\u2013315. \u2013 DOI: https:\/\/doi.org\/10.1007\/BF02823145<\/li>\n
  7. Chen Y. Reinforced evolutionary neural architecture search \/ Y. Chen, Q. Zhang, Ch. Huang, L. Mu, G. Meng, X. Wang \/\/ arXiv:1808.00193 [cs.NE]. \u2013 2018.<\/li>\n
  8. Litzinger S. Compute-efficient neural network architecture optimisation by a genetic algorithm \/ S. Litzinger, A. Klos, W. Schiffmann \/\/ Artificial Neural Networks and Machine Learning. \u2013 ICANN, 2019: Deep Learning. ICANN 2019. Lecture Notes in Computer Science, Springer, Cham, 2019. \u2013 Volume 11728. \u2013 DOI: https:\/\/doi.org\/10.1007\/978-3-030-30484-3_32<\/li>\n
  9. Costa M.G.F. Using convolutional neural networks with direct acyclic graph architecture in segmentation of breast lesions in US images \/ M.G.F. Costa, J.P.C. Mendes, W.C. A Pereira, C.F.F.C. Filho \/\/ VIII Latin American Conference on Biomedical Engineering and XLII National Conference on Biomedical Engineering. CLAIB 2019. IFMBE Proceedings, Springer, Cham, 2019. \u2013 Volume 75. \u2013 DOI: https:\/\/doi.org\/10.1007\/978-3-030-30648-9_99<\/li>\n
  10. Krizhevsky A. Learning multiple layers of features from tiny images \/ A. Krizhevsky \u2013 2009. \u2013 URL: https:\/\/www.cs.toronto.edu\/~kriz\/CIFAR-.html<\/li>\n
  11. Stamoulis D. Single-Path NAS: Designing hardware-efficient ConvNets in less than 4 hours \/ D. Stamoulis, R.\u00a0Ding, D. Wang, D. Lymberopoulos, B. Priyantha, J. Liu, D. Marculescu \/\/ arXiv:1904.02877 [cs.LG]. \u2013 2019.<\/li>\n
  12. Lozano M. Replacement strategies to maintain useful diversity in steady-state genetic algorithms \/ M. Lozano, F. Herrera, J.R. Cano \/\/ Advances in Soft Computing. \u2013 Springer, Berlin, Heidelberg, 2005. \u2013 Volume 32. \u2013 DOI: https:\/\/doi.org\/10.1007\/3-540-32400-3_7<\/li>\n
  13. Abadi M. TensorFlow: Large-scale machine learning on heterogeneous distributed systems \/ M. Abadi, A.\u00a0Agarwal, P. Barham. \/\/ OSDI’16 Proceedings of the 12th USENIX conference on Operating Systems Design and Implementation. \u2013 2016. \u2013 P. 265\u2013283. \u2013 ISBN: 978-1-931971-33-1.<\/li>\n
  14. Romanuke V.V. Appropriateness of DropOut layers and allocation of their 0.5 rates across convolutional neural networks for CIFAR-10, EEACL26, and NORB datasets \/ V.V. Romanuke \/\/ Applied Computer Systems. \u2013 2017. \u2013 Volume 22. \u2013 Issue 1. \u2013 P. 54\u201363. \u2013 DOI: https:\/\/doi.org\/10.1515\/acss-2017-0018<\/li>\n
  15. Radiuk P.M. Impact of training set batch size on the performance of convolutional neural networks for diverse datasets \/ P.M. Radiuk \/\/ Information Technology and Management Science. \u2013 2017. \u2013 Volume 20. \u2013 Issue 1. \u2013 P. 20\u201324. \u2013 DOI: https:\/\/doi.org\/10.1515\/itms-2017-0003<\/li>\n
  16. Meyes R. Ablation Studies in artificial neural networks \/ R. Meyes, L. Melanie, C.W. de Puiseau, T. Meisen \/\/ arXiv:1901.08644 [cs.NE]. \u2013 2019.<\/li>\n<\/ol>\n

     <\/p>\n","protected":false},"excerpt":{"rendered":"

    APPLICATION OF A GENETIC ALGORITHM TO SEARCH FOR THE OPTIMAL CONVOLUTIONAL NEURAL NETWORK ARCHITECTURE WITH WEIGHT DISTRIBUTION \u0417\u0410\u0421\u0422\u041e\u0421\u0423\u0412\u0410\u041d\u041d\u042f \u0413\u0415\u041d\u0415\u0422\u0418\u0427\u041d\u041e\u0413\u041e \u0410\u041b\u0413\u041e\u0420\u0418\u0422\u041c\u0423 \u0414\u041b\u042f \u041f\u041e\u0428\u0423\u041a\u0423 \u041e\u041f\u0422\u0418\u041c\u0410\u041b\u042c\u041d\u041e\u0407 \u0410\u0420\u0425\u0406\u0422\u0415\u041a\u0422\u0423\u0420\u0418 \u0417\u0413\u041e\u0420\u0422\u041a\u041e\u0412\u041e\u0407 \u041d\u0415\u0419\u0420\u041e\u041d\u041d\u041e\u0407 \u041c\u0415\u0420\u0415\u0416\u0406 \u0417 \u0420\u041e\u0417\u041f\u041e\u0414\u0406\u041b\u0415\u041d\u041d\u042f\u041c \u0412\u0410\u0413 \u0421\u0442\u043e\u0440\u0456\u043d\u043a\u0438: 7-11. \u041d\u043e\u043c\u0435\u0440: \u21161, 2020 (281) \u0410\u0432\u0442\u043e\u0440\u0438: P.M. RADIUK Khmelnytskyi National University \u041f.\u041c. \u0420\u0410\u0414\u042e\u041a \u0425\u043c\u0435\u043b\u044c\u043d\u0438\u0446\u044c\u043a\u0438\u0439 \u043d\u0430\u0446\u0456\u043e\u043d\u0430\u043b\u044c\u043d\u0438\u0439 \u0443\u043d\u0456\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442 DOI: https:\/\/www.doi.org\/10.31891\/2307-5732-2020-281-1-7-11 \u0420\u0435\u0446\u0435\u043d\u0437\u0456\u044f\/Peer review : 04. 01.2020 \u0440. […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[13],"tags":[],"_links":{"self":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/1145"}],"collection":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1145"}],"version-history":[{"count":3,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/1145\/revisions"}],"predecessor-version":[{"id":5066,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=\/wp\/v2\/posts\/1145\/revisions\/5066"}],"wp:attachment":[{"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1145"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1145"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/journals.khnu.km.ua\/vestnik\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1145"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}