Programming Languages Research Seminar
- ÕIS data: MTAT.03.271, 3 ECTS.
- Time: Thursdays 14:15-15:45.
- Place: Delta-2039 & Zoom (log into courses to see link).
- Language English.
- Contact: Vesal Vojdani.
This seminar is intended to introduce students to research conducted at the Laboratory for Software Science. This term, I will use the seminar to prepare and test out material for our upcoming course on program analysis, but here we will directly focus on three cutting-edge tools:
- CPAChecker embodies the state-of-the-art in software verification. It employs every trick in the toolbox from model checking and automated theorem proving to abstract interpretation in a configurable framework.
- ERAN: ETH Robustness Analyzer for Neural Networks is a sound static analyzer based on abstract interpretation that can be used for neural network certification. Our focus here, however, will be more on the underlying foundation of relational numeric abstract domains.
- Goblint is the world champion in static data race detection, having convincingly dominated the SVCOMP'21 race detection demo category. As impressive a feat as that may be, we are primarily interested in Goblint, and its younger sibling Põder, because they have by far the most welcoming developer communities.
The seminar will contain some introductory lectures by myself as well as student talks and/or a practical project; helping me preparing lab material for the upcoming course is always highly appreciated.
Requirements
- We expect basic familiarity and interest in the field of programming languages, but of course, we will answer any questions that participants have.
- Students must select topics directly related to the seminar focus.
- To pass the course, students need to give a single well-prepared presentation.