MAC-MERLin

Multi-Level Abstractions and
Causal Modeling for Enhanced
Reinforcement Learning

2024 – 2027

Integrating expert and domain knowledge into deep reinforcement learning agents through multi-level abstractions and causal modeling. We want to foster interpretable, more robust DRL agents that can generalize across environments.

Funding
Bundesministerium für Bildung und Forschung Grant 01IW24007

About

MAC-MERLin aims to integrate established expert and domain knowledge into artificial neural networks by leveraging multi-level abstractions in combination with causal modeling. Building on the established concepts of causality and abstraction, the project improves the interpretability of Deep Reinforcement Learning (DRL) agents and aligns them more closely with human understanding and expert insights, while at the same time making them more robust, better able to generalize across environments, and better tuned to real-world complexity.

The project addresses four fundamental challenges that currently make the practical deployment of neural-network decision making difficult: modeling, robustness of DRL agents, their ability to generalize to new environments, and their interpretability. By closing the gap between the theoretical and practical aspects of DRL, MAC-MERLin contributes to applications across smart cities and power grids, intelligent supply chains and production systems, autonomous driving, drug discovery, medical treatment and bioengineering.

Publications

2026

5 papers

Chaotic and complex dynamics expose the limits of counterfactual reasoning

2026 — Scientific Reports (Sci Rep)

@article{aalailaChaoticAndComplexDynamics2026,
 author = {Yahya Aalaila and Gerrit Großmann and Sumantrak Mukherjee and Jonas Wahl and Sebastian Vollmer},
 doi = {10.1038/s41598-026-52349-2},
 journal = {Scientific Reports (Sci Rep)},
 month = {6},
 number = {1},
 pages = {1-14},
 publisher = {Springer Nature},
 title = {Chaotic and complex dynamics expose the limits of counterfactual reasoning},
 url = {https://www.nature.com/articles/s41598-026-52349-2},
 volume = {16},
 year = {2026}
}
DOI URL

Per-Domain Generalizing Policies: On Learning Efficient and Robust Q-Value Functions

2026 — Proceedings of the Thirty-Sixth International Conference on Automated Planning and Scheduling. International Conference on Automated Planning and Scheduling (ICAPS-2026), June 27 - July 2, Dublin, Ireland

@inproceedings{mullerPerDomainGeneralizingPoliciesOn2026,
 author = {Müller, Nicola J. and Oster, Moritz and Valera, Isabel and Hoffmann, Jörg and Gros, Timo P.},
 booktitle = {Proceedings of the Thirty-Sixth International Conference on Automated Planning and Scheduling. International Conference on Automated Planning and Scheduling (ICAPS-2026), June 27 - July 2, Dublin, Ireland},
 doi = {10.1609/icaps.v36i1.42889},
 isbn = {ISSN 2334-0843},
 month = {6},
 number = {1},
 pages = {703-707},
 publisher = {AAAI Press},
 title = {Per-Domain Generalizing Policies: On Learning Efficient and Robust Q-Value Functions},
 url = {https://ojs.aaai.org/index.php/ICAPS/article/view/42889},
 volume = {36},
 year = {2026}
}
DOI URL

What are AI systems? Rethinking the core definition in the EU AI Act

2026 — Computer Law & Security Review

PDF
@article{meinelWhatAreAISystems2026,
 author = {Philip Meinel and Kevin Baum and Holger Hermanns and Emily Pöppelmann and Markus Langer and Anne Lauber-Ronsberg},
 doi = {10.1016/j.clsr.2026.106373},
 journal = {Computer Law & Security Review},
 number = {106373},
 pages = {1-12},
 publisher = {Elsevier},
 title = {What are AI systems? Rethinking the core definition in the EU AI Act},
 url = {https://www.sciencedirect.com/science/article/pii/S2212473X26001148},
 volume = {62},
 year = {2026}
}
DOI URL

PGTG: Procedurally Generated Grid-Based Traffic Gym

2026 — Reinforcement Learning Journal

PDF
@article{meyerPGTGProcedurallyGeneratedGridBased2026,
 author = {Joshua Meyer and Felix Maurice Kuntz and Verena Wolf and Jörg Hoffmann and Timo P. Gros},
 journal = {Reinforcement Learning Journal},
 pages = {},
 title = {PGTG: Procedurally Generated Grid-Based Traffic Gym},
 url = {https://rlj.cs.umass.edu/2026/papers/Paper32.pdf},
 volume = {7},
 year = {2026}
}
Code Poster

Toward a General Framework for Evaluating Per-Domain Generalization Using LLMs, Theorem Provers, and Statistical Model Checking

2026 — Workshop on Planning in the Era of LLMs at ICAPS-26

@inproceedings{mullerTowardAGeneralFramework2026,
 author = {Müller, Nicola J. and Rudolph, Naya and Taitler, Ayal and Gros, Timo P.},
 booktitle = {Workshop on Planning in the Era of LLMs at ICAPS-26},
 title = {Toward a General Framework for Evaluating Per-Domain Generalization Using LLMs, Theorem Provers, and Statistical Model Checking},
 url = {https://openreview.net/pdf?id=MLnzKKBdT4},
 year = {2026}
}
URL

2025

15 papers

Learning Per-Domain Generalizing Policies Using Offline Reinforcement Learning

2025 — Proceedings of the Ninth Workshop on Bridging the Gap Between AI Planning and (Reinforcement) Learning (PRL @ ICAPS 2025). Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning (PRL-2025), located at ICAPS'25, November 10, Melbourne, Australia

PDF
@inproceedings{mullerLearningPerDomainGeneralizingPolicies2025,
 author = {Nicola J. Müller and Moritz Oster and Timo P. Gros},
 booktitle = {Proceedings of the Ninth Workshop on Bridging the Gap Between AI Planning and (Reinforcement) Learning (PRL @ ICAPS 2025). Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning (PRL-2025), located at ICAPS'25, November 10, Melbourne, Australia},
 month = {11},
 publisher = {AAAI Press},
 title = {Learning Per-Domain Generalizing Policies Using Offline Reinforcement Learning},
 url = {https://prl-theworkshop.github.io/prl2025-icaps/papers/11.pdf},
 year = {2025}
}

PyDSMC: Statistical Model Checking for Neural Agents Using the Gymnasium Interface

2025 — Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems - Second International Joint Conference, Proceedings. International Joint Conference on Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems (QEST+FORMATS-2025), August 26-28, Aarhus, Denmark

@inproceedings{grosPyDSMCStatisticalModelChecking2025,
 author = {Timo P. Gros and Arnd Hartmanns and Ivo Hoese and Joshua Meyer and Nicola J. Müller and Verena Wolf},
 booktitle = {Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems - Second International Joint Conference, Proceedings. International Joint Conference on Quantitative Evaluation of Systems and Formal Modeling and Analysis of Timed Systems (QEST+FORMATS-2025), August 26-28, Aarhus, Denmark},
 doi = {10.1007/978-3-032-05792-1_8},
 editor = {Pavithra Prabhakar and Andrea Vandin},
 isbn = {978-3-032-05791-4},
 month = {10},
 pages = {134-156},
 publisher = {Springer Nature, Switzerland},
 series = {Lecture Notes in Computer Science, LNCS},
 title = {PyDSMC: Statistical Model Checking for Neural Agents Using the Gymnasium Interface},
 url = {https://link.springer.com/chapter/10.1007/978-3-032-05792-1_8},
 volume = {16143},
 year = {2025}
}
DOI URL

Visual Analysis of Action Policy Behavior: A Case Study in Grid-World Driving

2025 — World Conference on Explainable Artificial Intelligence. xAI: World Conference on Explainable Artificial Intelligence (xAI-2025), July 9-11, Istanbul, Turkey

@inproceedings{grosVisualAnalysisOfAction2025,
 author = {Timo P. Gros and David Groß and Julius Kamp and Stefan Gumhold and Jörg Hoffmann},
 booktitle = {World Conference on Explainable Artificial Intelligence. xAI: World Conference on Explainable Artificial Intelligence (xAI-2025), July 9-11, Istanbul, Turkey},
 month = {10},
 publisher = {Springer Berlin Heidelberg, Heidelberg},
 title = {Visual Analysis of Action Policy Behavior: A Case Study in Grid-World Driving},
 year = {2025}
}

Using Action-Policy Testing in RL to Reduce the Number of Bugs

2025 — Proceedings of the Eighteenth International Symposium on Combinatorial Search. International Symposium on Combinatorial Search (SoCS-2025), August 12-15, Scotland, United Kingdom

@inproceedings{eniserUsingActionPolicyTestingIn2025,
 author = {Hasan Ferit Eniser and Songtuan Lin and Nicola J. Müller and Anastasia Isychev and Valentin Wüstholz and Isabel Valera and Jörg Hoffmann and Maria Christakis},
 booktitle = {Proceedings of the Eighteenth International Symposium on Combinatorial Search. International Symposium on Combinatorial Search (SoCS-2025), August 12-15, Scotland, United Kingdom},
 doi = {10.1609/socs.v18i1.35990},
 editor = {Maxim Likhachev and Hana Rudová and Enrico Scala},
 isbn = {13 978-1-57735-901-2},
 month = {8},
 pages = {181-185},
 publisher = {AAAI Press, Washington, DC, USA},
 title = {Using Action-Policy Testing in RL to Reduce the Number of Bugs},
 url = {https://ojs.aaai.org/index.php/SOCS/article/view/35990},
 volume = {18},
 year = {2025}
}
DOI URL

Structural Causal Bottleneck Models

2025 — UAI 2025 Workshop on Causal Abstractions and Representations. Causal Abstractions and Representations (CAR), located at UAI 2025, July 25, Rio de Janeiro, Brazil

@inproceedings{bingStructuralCausalBottleneckModels2025,
 author = {Simon Bing and Jonas Wahl and Jakob Runge},
 booktitle = {UAI 2025 Workshop on Causal Abstractions and Representations. Causal Abstractions and Representations (CAR), located at UAI 2025, July 25, Rio de Janeiro, Brazil},
 month = {6},
 publisher = {OpenReview},
 title = {Structural Causal Bottleneck Models},
 url = {https://openreview.net/forum?id=EFNvKMSITB},
 year = {2025}
}
URL

A Roadmap for Responsible Robotics: Promoting Human Agency and Collaborative Efforts

2025 — IEEE Robotics & Automation Magazine

@article{araizaillanARoadmapForResponsible2025,
 author = {Dejanira Araiza-Illan and Kevin Baum and Helen Beebee and Raja Chatila and Sarah Moth-Lund Christensen and Simon Coghlan and Emily Collins and S.Kate Conroy and Alcino Cunha and Anna Dobrosovestnova and Hein Duijf and Vanessa Evers and Michael Fisher and Nico Hochgeschwender and Nadin Kökciyan and Séverin Lemaignan and Francisco Rodriguez-Lera and Sara Ljungblad and Martin Magnusson and Masoumeh Mansouri and Michael Milford and AJung Moon and Thomas M. Powers and Pericle Salvini and Teresa Scantamburlo and Nick Schuster and Marija Slavkovik and Ufuk Topcu and Daniel Vanegas and Andrzej Wasowski and Yi Yang},
 doi = {10.1109/MRA.2025.3620148},
 journal = {IEEE Robotics & Automation Magazine},
 number = {4},
 pages = {12-24},
 publisher = {IEEE},
 title = {A Roadmap for Responsible Robotics: Promoting Human Agency and Collaborative Efforts},
 url = {https://ieeexplore.ieee.org/abstract/document/11236462/authors#authors},
 volume = {32},
 year = {2025}
}
DOI URL

Disentangling AI Alignment: A Structured Taxonomy Beyond Safety and Ethics

2025 — AISoLA 2024. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2024)

@inproceedings{baumDisentanglingAIAlignmentA2025,
 author = {Kevin Baum},
 booktitle = {AISoLA 2024. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2024)},
 doi = {10.48550/arXiv.2506.06286},
 publisher = {Springer NAture Switzerland},
 title = {Disentangling AI Alignment: A Structured Taxonomy Beyond Safety and Ethics},
 url = {https://link.springer.com/chapter/10.1007/978-3-032-01377-4_8},
 year = {2025}
}
DOI URL

Justifications for Democratizing AI Alignment and Their Prospects

2025 — AISoLA 2025. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2025)

@inproceedings{baumJustificationsForDemocratizingAI2025,
 author = {Kevin Baum and Andre Steingrüber},
 booktitle = {AISoLA 2025. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2025)},
 publisher = {Springer Nature Switzerland},
 title = {Justifications for Democratizing AI Alignment and Their Prospects},
 url = {https://arxiv.org/abs/2507.19548},
 year = {2025}
}
URL

On the Complexities of Testing for Compliance with Human Oversight Requirements in AI Regulation

2025 — International Conference on Bridging the Gap between AI and Reality. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2025)

@inproceedings{langerOnTheComplexitiesOf2025,
 author = {Markus Langer and Veronika Lazar and Kevin Baum},
 booktitle = {International Conference on Bridging the Gap between AI and Reality. International Conference on Bridging the Gap between AI and Reality (AiSoLA-2025)},
 pages = {160-169},
 publisher = {Springer Nature Switzerland},
 title = {On the Complexities of Testing for Compliance with Human Oversight Requirements in AI Regulation},
 url = {https://www.dfki.de/fileadmin/user_upload/import/16858_978-3-032-07132-3.pdf},
 year = {2025}
}

How do we assess the trustworthiness of AI? Introducing the trustworthiness assessment model (TrAM)

2025 — Computers in Human Behavior

PDF
@article{schlickerHowDoWeAssess2025,
 author = {Nadine Schlicker and Kevin Baum and Alarith Ude and Sarah Sterz and Martin C. Hirsch and Markus Langer},
 editor = {Matthieu J. Guitton},
 journal = {Computers in Human Behavior},
 pages = {1-19},
 publisher = {Elsevier},
 title = {How do we assess the trustworthiness of AI? Introducing the trustworthiness assessment model (TrAM)},
 url = {https://www.sciencedirect.com/science/article/pii/S0747563225001189},
 volume = {170},
 year = {2025}
}
URL

The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applications

2025 — Proceedings of the Fourth Conference on Causal Learning and Reasoning. Conference on Causal Learning and Reasoning (CLeaR-2025)

PDF
@inproceedings{brouillardTheLandscapeOfCausal2025,
 author = {Philippe Brouillard and Chandler Squires and Jonas Wahl and Konrad Körding and Karen Sachs and Alexandre Drouin and Dhanya Sridhar},
 booktitle = {Proceedings of the Fourth Conference on Causal Learning and Reasoning. Conference on Causal Learning and Reasoning (CLeaR-2025)},
 organization = {PMLR},
 pages = {834-873},
 publisher = {MLResearchPress},
 title = {The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applications},
 url = {https://proceedings.mlr.press/v275/brouillard25a},
 volume = {275},
 year = {2025}
}
URL

Unitless Unrestricted Markov-Consistent SCM Generation: Better Benchmark Datasets for Causal Discovery

2025 — Proceedings of the Fourth Conference on Causal Learning and Reasoning. Conference on Causal Learning and Reasoning (CLeaR-2025)

PDF
@inproceedings{hermanUnitlessUnrestrictedMarkovConsistentSCM2025,
 author = {Rebecca J Herman and Jonas Wahl and Urmi Ninad and Jakob Runge},
 booktitle = {Proceedings of the Fourth Conference on Causal Learning and Reasoning. Conference on Causal Learning and Reasoning (CLeaR-2025)},
 pages = {1506-1531},
 publisher = {MLResearchPress},
 title = {Unitless Unrestricted Markov-Consistent SCM Generation: Better Benchmark Datasets for Causal Discovery},
 url = {https://proceedings.mlr.press/v275/herman25a.html},
 volume = {275},
 year = {2025}
}
URL

Separation-Based Distance Measures for Causal Graphs

2025 — Proceedings of The 28th International Conference on Artificial Intelligence and Statistics. International Conference on Artificial Intelligence and Statistics (AISTATS-2025), May 3-5, Mai Khao, Thailand

PDF
@inproceedings{wahlSeparationBasedDistanceMeasuresFor2025,
 author = {Jonas Wahl and Jakob Runge},
 booktitle = {Proceedings of The 28th International Conference on Artificial Intelligence and Statistics. International Conference on Artificial Intelligence and Statistics (AISTATS-2025), May 3-5, Mai Khao, Thailand},
 pages = {3412-3420},
 publisher = {Proceedings of Machine Learning Research},
 title = {Separation-Based Distance Measures for Causal Graphs},
 url = {https://proceedings.mlr.press/v258/wahl25b.html},
 volume = {258},
 year = {2025}
}
URL

Bridging the Gap Between AI Planning and Reinforcement Learning

2025 — AAAI 2025 Workshop Proceedings (PRL Workshop). Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning (PRL-2025), February 25

@inproceedings{ajanoviBridgingTheGapBetween2025,
 author = {Zlatan Ajanovi and Timo P. Gros and Floris Den Hengst and Daniel Holler and Harsha Kokel and Ayal Taitler},
 booktitle = {AAAI 2025 Workshop Proceedings (PRL Workshop). Workshop on Bridging the Gap Between AI Planning and Reinforcement Learning (PRL-2025), February 25},
 publisher = {AAAI Press},
 title = {Bridging the Gap Between AI Planning and Reinforcement Learning},
 url = {https://research.ibm.com/publications/bridging-the-gap-between-ai-planning-and-reinforcement-learning},
 year = {2025}
}
URL

Per-Domain Generalizing Policies: On Validation Instances and Scaling Behavior

2025 — The 35th International Conference on Automated Planning and Scheduling. International Conference on Automated Planning and Scheduling (ICAPS-2025), November 9-14, Melbourne, Australia

PDF
@inproceedings{grosPerDomainGeneralizingPoliciesOn2025,
 author = {Timo P. Gros and Nicola J. Müller and Daniel Fiser and Isabel Valera and Verena Wolf and Jörg Hoffmann},
 booktitle = {The 35th International Conference on Automated Planning and Scheduling. International Conference on Automated Planning and Scheduling (ICAPS-2025), November 9-14, Melbourne, Australia},
 publisher = {AAAI Press},
 title = {Per-Domain Generalizing Policies: On Validation Instances and Scaling Behavior},
 url = {https://doi.org/10.5281/zenodo.15314878},
 year = {2025}
}
URL

2024

1 paper

Motion Primitives as the Action Space of Deep Q-Learning for Planning in Autonomous Driving

2024 — IEEE Transactions on Intelligent Transportation Systems

@article{schneiderMotionPrimitivesAsThe2024,
 author = {Tristan Schneider and Matheus V. A. Pedrosa and Timo P. Gros and Verena Wolf and Kathrin Flaßkamp},
 doi = {10.1109/TITS.2024.3436530},
 journal = {IEEE Transactions on Intelligent Transportation Systems},
 month = {9},
 number = {11},
 pages = {17852-17864},
 publisher = {IEEE},
 title = {Motion Primitives as the Action Space of Deep Q-Learning for Planning in Autonomous Driving},
 volume = {25},
 year = {2024}
}
DOI