With the successful completion of project KP-06-N57/4, "Research and application of machine learning algorithms in the analysis and development of highly secure software," funded by the Scientific Research Fund, with project coordinator for Sofia University, prof. Milen Petrov (Faculty of mathematics and informatics), significant results were achieved in cybersecurity in modern digital environments. The project covers several key areas – the automation of the threat modeling process, the development of intelligent anomaly detection systems, and the use of artificial intelligence to protect network and IoT environments.
One of the main achievements is the creation of an intelligent system for protecting large volumes of data, which combines machine learning, threat recognition rules, genetic algorithms, and principal component analysis. The approach has been tested on the KDD Cup 1999 dataset, with results demonstrating high efficiency and the ability to recognize both known and new, unknown threats.
Another important area of research focuses on the use of large, open-source language models to simulate vulnerable IoT environments. These simulations have identified hidden weaknesses and tested responses to cyber incidents. A comparison between open local models and commercial cloud solutions highlights the advantages of local data processing and better control over sensitive information.
The project also investigated network threats, which pose a growing danger to modern information systems. A common architecture and prototype for protecting a network of devices has been developed, with data collected and analyzed in a graph database, allowing visualization of interrelationships and easier detection of suspicious activity. By applying predefined rules, restrictions are imposed on device behavior, preventing future malicious actions.
The results demonstrate that combining machine learning, intelligent analysis, and automated defense mechanisms is an effective approach to building more secure, sustainable, and reliable digital and IoT ecosystems.
For more information you can visit project web page: https://devops.tu-sofia.bg/ (with mirror on: https://devops2.w3c.fmi.uni-sofia.bg/).