Skip to main content


In most distributed and decentralized networks today, such as IoT and ad hoc networks, nodes are mobile due to the use of wireless connections. However, mobility brings new security issues. In dynamic environments, it becomes more complex to distinguish attacks from normal behaviour. On the one hand, attackers could easily evade security solutions in such networks. On the other hand, their effects in a dynamic network may be limited as the target nodes are also mobile. Machine learning-based solutions could better detect such complex behaviour in dynamically changing networks. In addition, transfer learning-based approaches could help networks adapt to new environments in a timely manner. In this talk, I will discuss not only mobility, but also other factors that can change the environment, such as the heterogeneity of nodes, the emergence of new attacks.