Перспективи побудови тестів контролю знань персоналу АЕС з використанням штучного інтелекту

• Видавець: State Scientific and Technical Center for Nuclear and Radiation Safety
• Дата: 2023
• Автори: Samoylov, V., Abramovych, R., Taranowski, А.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: State Scientific and Technical Center for Nuclear and Radiation Safety 2023
• Онлайн доступ: https://nuclear-journal.com/index.php/journal/article/view/1048

Although country and domain specific, the problem covered is still substantial and at least partially important beyond the nuclear energy sector of Ukraine. Among other key factors in the sphere of nuclear and radiation safety, personnel training requires that its effectiveness is maintained and improved on a constant basis. At some point in time, it appeared to be possible to raise the effectiveness of personnel training and knowledge assessment significantly through supplementing the applied scientific knowledge in pedagogy with one in computer science. Back then, a computer-based system for operating personnel training and attestation on the nuclear and radiation safety was developed within the cooperative programme supported by the European Union. The implementation thereof did allow transition at nuclear power plants from the classical training in classrooms to using the computer-based system in order for the personnel to do and complete training materials in stages, to assess their knowledge with testing and evaluate the results acquired. That did not cover, however, further automation of composing multi-choice tests, and this problem have remained unsolved, while the state of the art now does allow that to be done. Solving the problem covers considering the current approach to creation of knowledge assessment courses within the automated computer-based system for training and knowledge assessment titled ASKO, demonstrating its capacity to allow composing the multi-choice tests while using AI-related technologies, outlining available and ready to use AI-related technologies being relevant for that purpose, proposing the new approach that allows to compose the multi-choice tests within the considered computer-based system together with the outlined AI-related technologies, while maintaining the status quo with no additional costs. It is expected that raising the automation level of composing multi-choice tests under the proposed approach shall ensure improvement of personnel training effectiveness due to covering additional and substantially changed subject areas with knowledge assessment, making sure that the appropriate level of localisation for creating the knowledge assessment courses at the level of nuclear power plants is reached with the processes properly scaled, and having the related labour and non-labour resources optimised.