Our paper Constrained Training of Recurrent Neural Networks for Automata Learning has been presented at SEFM’22 in Berlin.

In this paper, written together with Bernhard K. Aichernig, Sandra König, Cristinel Mateis, Dominik Schmidt, and Martin Tappler, we propose a recurrent neural network architecture that enables the learning of a finite automata.

Our paper Active vs. Passive: A Comparison of Automata Learning Paradigms for Network Protocols has been presented at FMAS’22 in Berlin.

In this paper, written together with Bernhard K. Aichernig and Edi Muškardin, we compare active and passive automata learning techniques for learning network protocols, e.g., Bluetooth Low Energy or MQTT.