AI for Security



Polish computer scientist, Doctor of Economics and Lecturer at the Institute of Informatics, Faculty of Mathematics, Information Technology and Mechanics, University of Warsaw, Poland. Member of Ellis Society.

Amongst other things, we develop multi-level management systems to protect critical infrastructure as well as systems for securing key state services against both kinetic and cyber threats. Our algorithms develop concepts successfully used in the United States to improve the security of airports, sea terminals, and transmission infrastructure, or to increase the protection of endangered animal species in national parks in many places around the world. Furthermore, we build custom algorithmic solutions for crime detection and prevention. Finally, we develop AI-based solutions that increase the effectiveness of unmanned platforms in security-related applications.


In the last two decades, due to increased computing power, it has become possible to explicitly model many potential threats and possible countermeasures in security applications, especially those related to protecting critical infrastructure. To this end, a particularly popular approach is based on a model of Stackelberg games, where a defender of a critical infrastructure distributes limited security resources to guard a set of targets against different types of attackers and attacks. By applying novel optimisation methods,  it is possible to design defense strategies that significantly increase the chance of preventing a potential attack, which has been repeatedly shown in many practical applications. In particular, Stackelberg games were applied in such domains as critical infrastructure security (Los Angeles LAX Airport, Boston Harbour, New York), green security (natural parks and reserves in North and South America, Africa, and Asia), opportunistic crimes, as well as cybersecurity. Capitalising upon these experiences, our team builds advanced systems for protecting critical infrastructure in Poland.


Data science plays a crucial role in enhancing our ability to detect and prevent crime as well as in defense applications. While there exist a number of commercially available data science software solutions for the security sector, they typically do not include the latest scientific developments and particularly those based on AI technologies. This impedes the work of any analysts especially given the ever-increasing volumes of data. Hence, there is a growing need to build and deploy custom AI solution for the security sector that apply the newest scientific developments to large volumes of data in order to enhance the capabilities and efficiency of analysts.

Our team specialises in developing efficient algorithms for security-related applications with expertise ranging from graph algorithms to model and detect financial fraud to machine learning algorithms for crime analysis, forecasting, and prevention. For instance, by building upon well-established approaches in criminal science to select key variables, we can efficiently train relatively accurate machine learning models for predicting crimes in a given area. Then, by taking advantage of explainable machine learning techniques, we can discern the contributions of individual variables to the overall outcome to get a much deeper understanding of the roots and causes of particular types of crime.

We have pioneered an essential line of research into adversarial social network analysis which aims to understand how individuals and groups can evade social network analysis techniques, algorithms, and other tools. Under this paradigm, the strategic behaviour of network actors and network analysts is explicitly modeled. Such analysis has important practical implications. For instance, it may assist individuals and groups in protecting their privacy, by concealing their undisclosed relationship from link prediction algorithms etc. Furthermore, it may assist us in understanding the strategies employed by covert organisations to escape detection, and give rise to new strategy-proof countermeasures.

The research team “AI for Security” creates new solutions to strengthen private, public, and state security.


Our team also develops advanced solutions for various types of unmanned vehicles based on machine learning and other AI technologies. Recently, we have been particularly interested in multi-modal methods of communication between a human operator and an unmanned vehicle or a swarm of unmanned vehicles. A sample topic in this direction of research involves creating an AI-based planner which is able to understand and implement mission objectives from a plain-word description. We are proud to announce that some of our teammates involved in research on unmanned vehicles are participating in a prestigious international robotic competition: “The Mohamed Bin Zayed International Robotics Challenge” (as members of the Nomagic Warsaw MIMotaurs).




Tomasz P. Michalak is a Research Team Leader at the IDEAS NCBR research institute and a lecturer at the Institute of Informatics, Faculty of Mathematics, Informatics and Mechanics, University of Warsaw, Poland. He worked at the Department of Computer Science, University of Oxford, the School of Engineering and Computer Science, University of Southampton, the Department of Computer Science, University of Liverpool, and the Faculty of Applied Economics, University of Antwerp. He graduated from the Faculty of Economic Sciences, University of Warsaw, and holds a Ph.D. in economics from the Faculty of Applied Economics, University of Antwerp, Belgium. He has been a PI of two national and one international research projects. His results have been published by prestigious conferences and journals such as AAAI, IJCAI, IEEE ICDM, ACM EC, Nature Human Behaviour, Journal of Artificial Intelligence, and Games and Economic Behaviour.

Tomasz Michalak’s research interest focus on Artificial Intelligence, Social Networks, Computational Social Science, Multi-Agent Systems, Game Theory, Fintech,  E-Commerce. His specialty is the interface of Computer Science, Game Theory, and Economics, an area often called Algorithmic Game Theory. Recently, he has been especially interested in applications of cooperative and non-cooperative game theory to networks, social networks and blockchain, in particular.


Tomasz Michalak managed international and national research projects. Among other things, he held managerial roles in projects such as:

  • Project: “Tractable game-theoretic centrality measures”

Period: 2013-2017

Founder: National Science Centre, SONATA

  • Project: “Strategic Social Network Analysis”

Period: 2017-2022

Founder: National Science Centre, OPUS


Algorithms, especially the ones used in machine learning, promise to aid people in making decisions.

Blockchain technology was introduced in 2008. Several machine learning applications involve issues where privacy plays a special role.

Today, both science and industry rely heavily on machine learning models, predominantly artificial neural networks.

Skip to content