USING THE POSSIBILITIES OF A CONVULSIONAL NEURAL NETWORK TO DETECT POSITIVE ASSESSMENT OF DISCONTINUED STUDENTS
DOI:
https://doi.org/10.35546/kntu2078-4481.2024.2.35Keywords:
artificial intelligence, neural network, convolutional neural network, classification, intelligent systems, minimizing the risks of academic dishonesty.Abstract
Management of the quality of education is a key element of any modern educational system, which requires effective means of objective control of students’ educational achievements and exclusion of dishonesty. In Ukraine, a whole set of methods for evaluating educational success is actively being implemented in the practice of the educational process, both in the regular mode and in the remote mode. The topic of artificial intelligence, neural network training and research in this direction is an important criterion for measuring the technical level of research institutions, educational institutions or enterprises. The possibilities of using neural networks have not been fully explored. For many more years, they will be a means of information technology development and will require highly qualified IT specialists. The article reviews, systematizes, and summarizes publications on neural network training. It is suggested that with their help, facts of dishonesty can be detected when passing exams and assessments and minimize the risks of falsely determining the level of preparation of students. The methods of scientific research used in the work are: experiment, analysis of activity results. Among the theoretical research methods used: analysis, synthesis, comparison. The main results of the study. To conduct the experiment, a database of images of human behavior in a situation of stress and tension was created using the example of the intellectual role-playing game "Mafia" and using standard methods of the Keras library. Face selection is performed using the Viola-Jones method. The method uses sliding window technology. As a result of the experiment, images were selected with selected persons who did not behave virtuously when passing the exam. The accuracy is quite high, but errors are possible. Scientific novelty. In order to detect fraud in the educational field when taking tests and exams, it is proposed to use the capabilities of a convolutional neural network, the work of which will be aimed at classifying images with respect to integrity. To detect fraud, when determining the level of preparation of students, the following algorithm was used: 1. Converting the frame to a black and white image. 2. Selecting a face for analysis. 3. Image preparation for neural network processing. 4. Classification of student behavior.
References
Ioffe, S., Szegedy, C.: “Batch normalization: Accelerating deep network training by re-ducing internal covariate shift,”arXiv preprint arXiv:1502.03167, Feb. (2015). URL: https://arxiv.org/abs/1502.03167/ (дата звернення: 01.02.2024).
Xiao, T., Zhang, J.: et al., “Error-driven incremental learning in deep convolutional neu-ral network for large-scale image classification,” in International Conference on Multi-media, no. 22. ACM, pp. 177–186 (2014)
Тымчук, А.: «Метод распознавания лиц Виолы-Джонса (Viola-Jones).(2015). URL: https://oxozle.com/2015/04/11/metod-raspoznavaniya-lic-violy-dzhonsa-viola-jones/ (дата звернення: 01.02.2024).
Baskin, C., Natan, L., Avi, M., Zheltonozhskii, E. (2017). Streaming Architecture for Large-Scale Quantized Neural Networks on an FPGA-Based Dataflow Platform.
Max pooling / pooling. (2018). Режим доступу: https://computersciencewiki.org/index.php/Max-pooling_/_Pooling/ (дата звернення: 11.02.2024).
Alom, Md., Z., Taha., T., Yakopcic, C., Westberg, S., Sidike, P., Nasrin., M., Hasan, M., Essen, B., Awwal, A., Asari, V.: A State-of-the-Art Survey on Deep Learning Theo-ry and Architectures. (2019). Electronics. 8. 292. 10.3390/electronics8030292.
Gibson, A., Patterson, J.: Deep Learning. O’Reilly Media, Inc., (2017). URL: https://www.safaribooksonline.com/library/view/deep-learning/9781491924570/(дата звернення: 11.02.2024).
Fishman T. (2012). The Fundamental Values of Academic Integrity (2nd edition). In-ternational Center for Academic Integrity, Clemson University. URL: http://www.academicintegrity.org/icai/assets/AUD_Integrity_Quotes.pdf
