The new standalone program detects viruses by itself .
Russian programmers have created an AI-based antivirus that can independently detect and neutralize various types of malware without the need for an Internet connection. This is reported by Izvestia, referring to the descriptions of the product developers.
The developers of the new antivirus are specialists from Penza State University. According to students, the software is based on machine learning and does not require updates. The program is already ready for use on Windows and is able to detect malware types such as Trojans, rootkits, and software that illegally uses computer resources. Project leader Daniil Mishin emphasized that the key difference between the product is that there is no need for a permanent Internet connection.
A neural network written in Python analyzes the behavior of software on your computer and detects signs of malicious activity. To improve the accuracy of detecting malware, the program uses two types of analysis at once — neural network and neurosignature:
In 2024, the team plans to complete the development of the program and start its certification in November. The antivirus will be available to both individuals and companies with the planned subscription distribution model.
Industry experts believe that such a promising direction can significantly strengthen the protection of computers, but they believe that this technology should complement, and not completely replace, existing solutions.
Representatives of the IT and information security industry note that the use of neural network technologies and machine learning algorithms is a promising direction in cybersecurity. Such methods allow detecting malware in 96% of cases, which is an excellent result. However, antivirus programs cannot do without regular updates, even if they are based on machine learning, experts say.
It is noteworthy that the principle underlying the new antivirus is shared by many IT companies. It is expected that in the future such products can not only improve, but also replace existing solutions in the field of information security.
Russian programmers have created an AI-based antivirus that can independently detect and neutralize various types of malware without the need for an Internet connection. This is reported by Izvestia, referring to the descriptions of the product developers.
The developers of the new antivirus are specialists from Penza State University. According to students, the software is based on machine learning and does not require updates. The program is already ready for use on Windows and is able to detect malware types such as Trojans, rootkits, and software that illegally uses computer resources. Project leader Daniil Mishin emphasized that the key difference between the product is that there is no need for a permanent Internet connection.
A neural network written in Python analyzes the behavior of software on your computer and detects signs of malicious activity. To improve the accuracy of detecting malware, the program uses two types of analysis at once — neural network and neurosignature:
- Neural network analysis is based on machine learning — AI already has algorithms for the behavior of software, such as Trojan viruses. For example, if a program detects that certain software is being copied spontaneously to various folders, it is classified as a virus and further distribution is prevented.
- Neurosignature analysis combines traditional malware detection techniques with artificial intelligence. The developers explain that it is difficult to create a malicious file that is completely different from its predecessors, so the AI constantly analyzes their behavior.
In 2024, the team plans to complete the development of the program and start its certification in November. The antivirus will be available to both individuals and companies with the planned subscription distribution model.
Industry experts believe that such a promising direction can significantly strengthen the protection of computers, but they believe that this technology should complement, and not completely replace, existing solutions.
Representatives of the IT and information security industry note that the use of neural network technologies and machine learning algorithms is a promising direction in cybersecurity. Such methods allow detecting malware in 96% of cases, which is an excellent result. However, antivirus programs cannot do without regular updates, even if they are based on machine learning, experts say.
It is noteworthy that the principle underlying the new antivirus is shared by many IT companies. It is expected that in the future such products can not only improve, but also replace existing solutions in the field of information security.
