The first phase of the joint research project of Quadron Cyber Defense Kft., BME and the National Artificial Intelligence Laboratory (MILab) has been completed. At the center of the research is an algorithm capable of using artificial intelligence to detect unusual IT system behavior that could be indicative of attacks, long before damage occurs. on a previous BME collaboration that began under the university’s PARIPA program, during which talented university students are given the opportunity to participate in research projects of the Quadron research laboratory. internal security threats arising in computer networks in real time. Research continued within MILab with the collaboration of Quadron and BME.
“Anomaly detection aims to detect the operation of a controlled system that deviates from normal behavior, which can be done using algorithms suitable for detecting “unusual” data. Network anomalies can be detected by various causes such as failure of network components, viruses and cyber attacks. Timely detection of such events is critical to ensure the reliability, availability, and quality of the network and related services. With the rapid growth in the number of new devices connected, the launch of new services and the use of new applications, this is constantly becoming more of a challenge,” said Balazs Sekeres, CTO of Quadron.
“MILab aims to highlight cutting-edge scientific and engineering competencies of Hungary, as well as to ensure the effective transfer of knowledge in the field of artificial intelligence. Over the past year, the Budapest University of Technology conducted research and development within the framework of MILab, aimed, among other things, at the security and protection of personal data, machine vision and sensors, intelligent manufacturing based on machine learning, applications for medicine and healthcare. ”, said Dr. Janosz Lewendowski, BME Academic Vice-Chancellor, University Professor, MILAB BME Coordinator.
The procedure has been implemented and tested as a so-called NFPlugin extension, which is a component of the NFStream Framework. The open source platform was created in collaboration with Dr. Zied Aouini, SoftAtHome researcher in France, and Dr. Adrian Pekar, Associate Professor at BME HIT. The purpose of their collaboration is to help and popularize network data traffic research through a structure accessible to all.
“NFStream is a Python framework that provides fast, flexible, and expressive data structures. enable online and make working with offline network data intuitive and easy, and facilitate the transition from network to machine learning. which helps intelligent network management aimed at ensuring the quality and reliability of Internet applications and services.” – Dr. Adrian Pekar reported on the collaboration.
After the first phase of the study was completed, the Quadron R&D lab went ahead with the new behavior-based process so that it could become a useful and implementable solution for its partners in the shortest possible time. The work is already in the development of a prototype, in which BME students will continue to participate, who, thanks to this, will be able to gain a unique professional experience. The models have great potential. future scientific research. Our cooperation has not only research, but also market potential, such as the inclusion of a procedure in the security system of the Quadron Data Lake cyber defense center. can provide significant unique value-added service content to our domestic and international customers, thereby further increasing our service exports,” said Peter Gulyas, Managing Director of Quadron. were also found that could lay the foundation for a long-term collaboration between Quadron Research Lab, MILAB, and BME to continue applied research.
Hardware, software, tests, interesting and colorful news from the IT world by clicking here!
