Simon Kohaut

PhD Candidate · AI & ML Lab · TU Darmstadt

profile.webp

I am a PhD candidate at the Artificial Intelligence & Machine Learning Lab at TU Darmstadt, co-supervised by Kristian Kersting and Devendra Singh Dhami.

My research sits at the intersection of neuro-symbolic AI and responsible reasoning systems, developing methods that combine the representational powers of neural learning with the interpretability and reasoning capabilities of symbolic systems. I am particularly interested in deploying such systems in safety-critical domains, including robotics and autonomous transportation, where reliability, explainability, and formal guarantees are essential.

I care about building AI that is not only performant, but also verifiable, transparent, and safe by design, contributing to a future where intelligent systems can be responsibly integrated into the physical world.

news

Jun 11, 2026 Paper accepted at Data-centric Machine Learning Research (DMLR): CycliST: A video language model benchmark for reasoning on cyclical state transitions.
Mar 03, 2026 New preprint: Right in time: Reactive reasoning in regulated traffic spaces.
Feb 07, 2026 New preprint: Reactive knowledge representation and asynchronous reasoning.
Nov 29, 2025 New preprint: CycliST: A video language model benchmark for reasoning on cyclical state transitions
Oct 18, 2025 Paper published at 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS): The Constitutional Filter: Bayesian estimation of compliant agents.
Oct 02, 2025 Journal paper accepted in IEEE Transactions on Intelligent Transportation Systems: Probabilistic mission design for neuro-symbolic unmanned aircraft systems.
Sep 24, 2025 Journal paper published in ACM Journal on Autonomous Transportation Systems: Hybrid many-objective optimization in probabilistic mission design for compliant and effective UAV routing.
Jul 20, 2025 New preprint: The Constitutional Controller: Doubt-calibrated steering of compliant agents.
Dec 18, 2024 New preprint: Answer set networks: Casting answer set programming into deep learning.
Aug 31, 2024 Paper published at 2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC): StaR Maps: Unveiling uncertainty in geospatial relations.
Jun 03, 2024 Paper published at 2024 International Conference on Unmanned Aircraft Systems (ICUAS): Towards probabilistic clearance, explanation and optimization.
Oct 25, 2023 US Patent filed: Probabilistic mission design (App. 18/192,660).
Sep 03, 2023 Paper published at 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC): Mission design for unmanned aerial vehicles using hybrid probabilistic logic programs.
Sep 30, 2022 I started my PhD at the AI & ML Lab at TU Darmstadt in collaboration with the Honda Research Institute Europe.
Sep 23, 2021 Paper published at 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC): Error decomposition for hybrid localization systems.
May 26, 2021 US Patent filed: Analysis of localization errors in a mobile object (App. 17/105,510).
Jun 30, 2018 Paper published at International Robotic Sailing Conference 2018 (IRSC): Dynamic simulation model for an autonomous sailboat.

selected publications

  1. ACM JATS
    Hybrid many-objective optimization in probabilistic mission design for compliant and effective UAV routing
    Simon Kohaut, Nikolas Hohmann, Sebastian Brulin, Benedict Flade, Julian Eggert, Markus Olhofer, Jürgen Adamy, Devendra Dhami, and Kristian Kersting
    Journal on Autonomous Transportation Systems, 2025
  2. IEEE/RSJ IROS
    The Constitutional Filter: Bayesian Estimation of Compliant Agents
    Simon Kohaut, Felix Divo, Benedict Flade, Devendra Singh Dhami, Julian Eggert, and Kristian Kersting
    In 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
  3. IEEE T-ITS
    Probabilistic mission design for neuro-symbolic unmanned aircraft systems
    Simon Kohaut, Benedict Flade, Daniel Ochs, Devendra Singh Dhami, Julian Eggert, and Kristian Kersting
    IEEE Transactions on Intelligent Transportation Systems, 2025
  4. DMLR
    CycliST: A Video Language Model Benchmark for Reasoning on Cyclical State Transitions
    Simon Kohaut, Daniel Ochs, Shun Zhang, Benedict Flade, Julian Eggert, Kristian Kersting, and Devendra Singh Dhami
    Journal of Data-centric Machine Learning Research, 2026
  5. arXiv
    Reactive Knowledge Representation and Asynchronous Reasoning
    Simon Kohaut, Benedict Flade, Julian Eggert, Kristian Kersting, and Devendra Singh Dhami
    arXiv preprint arXiv:2602.05625, 2026