METHODS OF MACHINE LEARNING IN MODERN METROLOGY

Автори:
1
Kharkiv National University of Radio Electronics, Ukraine

In the modern world of scientific and technological progress, the requirements for the accuracy and reliability of measurements are becoming increasingly stringent. The rapid development of machine learning (ML) methods opens up perspectives for improving metrological processes and enhancing the quality of measurements. This article explores the potential application of ML methods in metrology, outlining the main types of ML models in automatic instrument calibration, analysis, and prediction of data. Attention is paid to the development of hybrid approaches that combine ML methods with traditional metrological methods for the optimal solution of complex measurement tasks.

  1. S. Artemuk, I. Mykytyn, A.O. Kozhevnikov. "Investigation of methods for determining the accuracy of metrological measurements",   2022.  [Online]. Available:https://openarchive.nure.ua/entities/publication/6c386f59- 2585-4e39-82a0-f452a70a8494
  2. A. Mueller, S. Guido. "Introduction to Machine Learning with      Python",  2016-2017. [Online]. Available: https://library-it.com/wp-content/uploads/2021/01/a_ myuller_s_gvido_vvedenie_ v_mashinnoe.pdf
  3. Y.O. Hryshkun, S.M. Kravchenko, A.Yu. Levchenko, Yu.I. Lysogor. "Machine Learning Methods" in Znanstvena misel journal, No. 39, 2020, p. 55. [Online]. Available: https://www.znanstvena-journal.com/wp- content/uploads/2020/10/Znanstvena-misel-journal-%E2%84%9639-2020-VOL.1.pdf#page=55
  4. A.M. Rasulova, A.V. Izmajlova. "Application of the Isola- tion Forest algorithm for justifying the uniqueness of water bodies in a group of karst lakes" in Earth sciences, 2021. doi:10.33619/2414-2948/72/08
  5. S. Bilson, A. Thompson, D. Tucker, J. Pierce «A machine learning approach to automation and uncertainty evaluation for self-validating thermocouples» in NIST.SP.2100-05, Di- gest Conference ITS10, 2023.
  6. J.V. Pierce, O. Ongrey, G. Machin, S.D. Sweeney. "Self- Validating Thermocouples Based on Fixed Points at High Temperatures" in Metrologia, Vol. 47, No. 1, 2010. doi:10.1088/0026-1394/47/1/L01
  7. M.Y. Krajnyuk. "Deep Learning for Measurements: Appli- cation of Neural Networks in Solving Real Problems" in Metrological Aspects of Decision Making in the Conditions of Work at Technogenic Hazardous Objects. [Online]. Available: https://er.chdtu.edu.ua/bitstream/ChSTU/ 4662/1/sbornik_konf_2023.pdf#page=21
  8. O.M. Vasilevsky. "Concept of Metrological Support in Industry 4.0" in Information Technologies and Computer Engineering, Vol. 48, No. 2, 2020. doi:10.31649/1999-9941-2020-48-2-37-44
  9. Padraig Timoney et al. "Implementation of machine learn- ing for high-volume manufacturing metrology challenges" in Proceedings, Volume 10585, Metrology, Inspection, and Process Control for Microlithography XXXII; 105850X, 2018. doi:10.1117/12.2300167
  10. Marcela Vallejo, Carolina de la Espriella, Juliana Gómez- Santamaría, Andrés Felipe Ramírez-Barrera, and Edilson Delgado-Trejos. "Soft metrology based on machine learn- ing: a review" in Measurement Science and Technology, Vol. 31, No. 3, 2019. doi:10.1088/1361-6501/ab4b39