Infections as Abstract Symbolic Finite Automata: Formal Model and Applications

In this paper, we propose a methodology, based on machine learning, for building a symbolic finite state automata-based model of infected systems, that expresses the interaction between the malware and the environment by combining in the same model the code and the semantics of a system and allowing to tune both the system and the malware code observation. Moreover, we show that this methodology may have several applications in the context of malware detection.