Event Calendar
Computer Science Seminar | Timed Automata and How to Learn Them
- This event has passed.
.png)
Dear All,
The Department of Computer Science invites you to a seminar on "Timed Automata and How to Learn Them"
Zoom Link: https://zoom.us/j/98106130931?pwd=82IYpMhDN3urjrv0Sw7yTbh86XMZaZ.1
Abstract: Automata are fundamental models in verification, testing, and synthesis, but constructing them manually is often difficult and error-prone. Automata learning is an automated synthesis technique that aims at synthesizing automata models, that are correct-by-construction, from queries or observations. It has been successfully applied to a wide range of applications such as verification of network protocols, reverse engineering legacy systems etc. Many systems, however, depend not only on the order of events but also on the timing of their occurrences. Timed automata capture this behaviour by extending finite automata with clocks. However, learning them is challenging due to the continuous nature of time. In this talk, I will present our works on passive and active learning of timed languages recognizable by Event-Recording Automata (ERA), a subclass of Timed Automata. In the passive learning framework, we provide a state-merging algorithm to construct an ERA from positive and negative data. In the active setting, we propose a novel algorithm to be able to learn an ERA with the minimum number of states that accepts the target language.
This is based on joint work with Sayan Mukherjee and Jean-François Raskin.
About the Speaker: Anirban Majumdar is currently Prof. R. Narasimhan Postdoctoral Fellow in the Computer Science Department at the Tata Institute of Fundamental Research, working with Shibashis Guha. Earlier he was a post-doctoral researcher in the Verification group at Université Libre de Bruxelles, working with Jean-François Raskin. He obtained his PhD from Université Paris-Saclay under the supervision of Patricia Bouyer-Decitre and Nathalie Bertrand. His research interests include automata learning, controller synthesis for reactive systems, and synthesis under uncertainty.
We look forward to your active participation.