Choose a city

Neural network taught to catch cheating students

Neural network taught to catch cheating students

In the Republic, an innovative system was invented that provides surveillance of students who pass the exam. It records all suspicious actions and transmits data about them to the teacher. This development was created by two students and graduates of Nazarbayev University - Miras Sovetov and Margulan Yermek.

Systems that allow you to take the exam online have become very popular during the coronavirus pandemic. Now they are used in a huge number of countries around the world. Online exams are taken by schoolchildren, College and University students, as well as candidates for positions in large companies.

Graduates of the NU School of engineering and digital Sciences have created a program that runs on the basis of artificial intelligence. It is able to recognize the face of the exam taker, as well as their speech and any speech that sounds close to the person taking the exam. During the exam, the system monitors whether the student turns away, looks around, whether there is a substitution by another person or help from a third person. The system also tracks the activity of the browser and other applications. When you try to open another browser, the app takes a screenshot of the student's screen.

According to the developers, the system does not decide who wrote off and who did not. The neural network only records all suspicious moments and sends the teacher a full report on each of them.

Only six people were involved in the development of the project, and it took them about six months. The program was decided to be called AERO. The developers of the system note that it has already made the work of teachers much easier in several universities. Moreover, as the creators of the neural network note, it does not need powerful computers for full operation, it does not use a lot of computer resources and can work at full capacity even on a smartphone.

Write a review

Note: HTML is not translated!
    Bad           Good