Using ai to detect indicators of electoral fraud
Автор: Kilyachkov Anatoliy Anatolevich, Chaldaeva Larisa Alekseevna, Korolev Denis Andreevich, Bayer Andrey Viktorovich
Журнал: Власть @vlast
Рубрика: Экспертиза
Статья в выпуске: 5, 2021 года.
Бесплатный доступ
In the modern world, one of the ways for Russia to achieve a leading position is to have better quality of management decisions. This requires new methods and digital technologies of management, including the use of artificial intelligence or (in more correct and neutral terms) neural networks. One can solve the task using neural networks if it requires an intellectual effort and is already performed by a person. In addition, there must be a large database with information on the problem. Identification of fraud during the elections is one of the important problems that can be solved using neural networks. Use of neural networks is possible as the election process at polling stations is monitored by video surveillance, where the video recordings are checked by controllers. We estimated that the 2021 elections would record 7.5 million hours, which is sufficient to develop and train the neural networks. The article discusses how the information recorded during the elections can help to fraud identification. The author notes that neural networks should be used in the recommendations capacity, leaving it to the individuals to make the final decision.
Artificial intelligence, neural networks, elections, identification of facts of fraudulent activities, video surveillance
Короткий адрес: https://sciup.org/170191668
IDR: 170191668 | DOI: 10.31171/vlast.v29i5.8546