Seminar in Theoretical CS, Summer 2026
News
- 10.12.2025: we are online
Dates & Deadlines
| TBA | Kick-off meeting (room 4201b, 2nd floor, building E1, Faculty of CS) |
| TBA | Topic preferences due |
| TBA | Detailed outline due |
| TBA | Seminar report due |
| TBA | Presentation slides due |
| TBA | Presentations (room 4201b, 2nd floor, building E1, CS Department) |
Note that the full versions of your report and your slides should be your final submission and the camera-ready versions should differ only with regard to minor remarks, comments, and corrections by your supervisor. Please feel free, however, to talk to your supervisor about submitting preliminary versions before the due dates.
Introduction
- Slides of introduction (to follow)
- Topic choice form (to follow)
Overview
The success of machine learning has been motivating researchers in formal methods to adapt learning methods to the verification setting, where correctness guarantees on the result are essential. The aim of this seminar is to provide an overview of approaches to exploit learning methods in verification or to verify systems with learning-based components. This includes, for instance,
- the use of learning techniques (e.g., reinforcement learning) for speeding up verification,
- the use of machine learning data structures and algorithms (e.g. decision trees) for enhancing results of verification (e.g., by generating simple invariants of programs or by generating efficient controllers for reactive systems),
- verification of machine-learning artefacts (e.g. verification of neural networks), or
- the meta-usage of machine learning (e.g. to predict the best tools to be applied to a verification problem).
The setting of the seminar is inspired by an annual workshop of the same name. It covers several areas of computer science which combine learning and verification topics, and which are closely connected to the research fields of the respective supervisors. Each area features a number of topics which are discussed in a scientific journal or conference article. These research articles are the basis on which students have to prepare their report and presentation. The following list gives an overview of the areas that will be covered.
Topic Areas (preliminary & incomplete)
A. Verification of Neural Networks [Xaver Fink]
- An Abstract Domain for Certifying Neural Networks
- Paper: Gagandeep Singh, Timon Gehr, Markus Püschel, Martin Vechev: An Abstract Domain for Certifying Neural Networks. POPL 2019
- Student: –
- Supervisor: –
- Solving Probabilistic Verification Problems of Neural Networks
- Paper: David Boetius, Stefan Leue, Tobias Sutter: Solving Probabilistic Verification Problems of Neural Networks using Branch and Bound. ICML 2025
- Student: –
- Supervisor: –
B. Generation of Invariants [Darion Haase]
- Neural Model Checking
- Mirco Giacobbe, Daniel Kroening, Abhinandan Pal, Michael Tautschnig: Let a Neural Network be Your Invariant, NeurIPS 2025
- Student: –
- Supervisor: –
C. Controller Synthesis for Partially Observable Stochastic Systems [Alexander Bork, Lisa Pühl]
- Strategy Synthesis for Partially Observable Markov Decision Processes
- Steven Carr, Nils Jansen, Ralf Wimmer, Alexandru C. Serban, Bernd Becker, Ufuk Topcu: Counterexample-Guided Strategy Improvement for POMDPs Using Recurrent Neural Networks, arXiv 2019
- Student: –
- Supervisor: –
D. Compositional Learning [Hannah Mertens]
- Compositional Learning
- Léo Henry, Mohammad Reza Mousavi, Thomas Neele, Matteo Sammartino: Compositional Active Learning of Synchronizing Systems Through Automated Alphabet Refinement, CONCUR 2025
- Student: –
- Supervisor: –
E. Learning for Software Verification [Thomas Noll]
- Detecting and Locating Null Pointer Vulnerabilities
- Yue Ding, Qian Wu, Yinzhu Li, Dongdong Wang, Jiaxin Huang: Leveraging Deep Learning Models for Cross-function Null Pointer Risks Detection, AITest 2023
- Student: –
- Supervisor: –
- Bisimulation Learning
- Alessandro Abate, Mirco Giacobbe, Yannik Schnitzer: Bisimulation Learning, CAV 2024
- Student: –
- Supervisor: –
Prerequisites
Basic knowledge in the following areas is expected:
- Formal languages and automata theory
- Mathematical logic
Previous knowledge in neural networks, stochastic systems and quantum programming (depending on the choice of the topic) is helpful but not mandatory.
Registration
Registration to the seminar is handled via the SuPra system.
Additional Resources
- Report template (to follow)
- Presentation template (to follow)
- Grading scheme
- How to Write a Seminar Paper
- Ethische Richtlinien für das Verfassen wissenschaftlicher Arbeiten
- How to Give Presentations
- Introduction to LaTeX