@article{AmmelingGanzRosbachetal.2026, author = {Ammeling, Jonas and Ganz, Jonathan and Rosbach, Emely and Lausser, Ludwig and Bertram, Christof and Breininger, Katharina and Aubreville, Marc}, title = {Benchmarking Foundation Models for Mitotic Figure Classification}, volume = {3}, pages = {2026:003}, journal = {Machine Learning for Biomedical Imaging}, number = {MELBA-BVM 2025 Special Issue}, publisher = {Melba editors}, address = {[s. l.]}, issn = {2766-905X}, doi = {https://doi.org/10.59275/j.melba.2026-a3eb}, pages = {38 -- 55}, year = {2026}, abstract = {The performance of deep learning models is known to scale with data quantity and diversity. In pathology, as in many other medical imaging domains, the availability of labeled images for a specific task is often limited. Self-supervised learning techniques have enabled the use of vast amounts of unlabeled data to train large-scale neural networks, i.e., foundation models, that can address the limited data problem by providing semantically rich feature vectors that can generalize well to new tasks with minimal training effort increasing model performance and robustness. In this work, we investigate the use of foundation models for mitotic figure classification. The mitotic count, which can be derived from this classification task, is an independent prognostic marker for specific tumors and part of certain tumor grading systems. In particular, we investigate the data scaling laws on multiple current foundation models and evaluate their robustness to unseen tumor domains. Next to the commonly used linear probing paradigm, we also adapt the models using low-rank adaptation (LoRA) of their attention mechanisms. We compare all models against end-to-end-trained baselines, both CNNs and Vision Transformers. Our results demonstrate that LoRA-adapted foundation models provide superior performance to those adapted with standard linear probing, reaching performance levels close to 100 \% data availability with only 10 \% of training data. Furthermore, LoRA-adaptation of the most recent foundation models almost closes the out-of-domain performance gap when evaluated on unseen tumor domains. However, full fine-tuning of traditional architectures still yields competitive performance.}, language = {en} } @article{MohanSchoen2026, author = {Mohan, Adithya and Sch{\"o}n, Torsten}, title = {Toward Robust Agents: A Survey of Adversarial Attacks and Defenses in Deep Reinforcement Learning}, volume = {14}, journal = {IEEE Access}, publisher = {IEEE}, address = {New York}, issn = {2169-3536}, doi = {https://doi.org/10.1109/ACCESS.2026.3657855}, pages = {14481 -- 14497}, year = {2026}, abstract = {Deep Reinforcement Learning (DRL) has demonstrated remarkable success in autonomous decision-making across diverse domains, including robotics, autonomous driving, and game playing. However, recent studies have uncovered a critical vulnerability: DRL agents are highly susceptible to adversarial attacks that can significantly degrade their performance or lead to catastrophic failure. These attacks exploit different components of the learning pipeline observations, actions, rewards, and policies exposing new challenges unique to DRL compared to supervised learning. This survey provides a comprehensive examination of adversarial threats and corresponding defense mechanisms within the DRL paradigm. This also aims to serve as a foundational reference for researchers and practitioners seeking to understand and mitigate adversarial vulnerabilities in DRL.}, language = {en} } @inbook{SprogiesSchmidt2015, author = {Sprogies, Marc and Schmidt, Werner}, title = {Introducing S-BPM at an IT Service Providers}, booktitle = {S-BPM in the Wild: Practical Value Creation}, editor = {Fleischmann, Albert and Schmidt, Werner and Stary, Christian}, publisher = {Springer}, address = {Cham}, isbn = {978-3-319-17542-3}, doi = {https://doi.org/10.1007/978-3-319-17542-3_4}, pages = {55 -- 74}, year = {2015}, abstract = {IT Service consumers have a clear idea of agile, flexible and transparent service processes to quickly get their needs satisfied. For an IT service provider like WK EDV GmbH this arises the challenge of designing its procedures adequately. For that reason WK decided to consolidate and optimize their service processes. It ran a pilot project to analyze, redesign and newly implement the software deployment process which is part of their overall Application Lifecycle Management (ALM) process. The project team applied Subject-oriented Business Process Management (S-BPM) as methodology and the Metasonic Suite as the respective software toolset in order to gain insights into and experience with the S-BPM environment. This contribution reports on the course of the project, the results and the learnings.}, language = {en} } @unpublished{SalamatMatternOlzemetal.2025, author = {Salamat, Babak and Mattern, Dominik and Olzem, Sebastian-Sven and Elsbacher, Gerhard and Seidel, Christian and Tonello, Andrea M.}, title = {GMP3: Learning-Driven, Bellman-Guided Trajectory Planning for UAVs in Real-Time on SE(3)}, publisher = {arXiv}, address = {Ithaca}, doi = {https://doi.org/10.48550/arXiv.2509.21264}, year = {2025}, abstract = {We propose GMP3, a multiphase global path planning framework that generates dynamically feasible three-dimensional trajectories for unmanned aerial vehicles (UAVs) operating in cluttered environments. The framework extends traditional path planning from Euclidean position spaces to the Lie group SE(3), allowing joint learning of translational motion and rotational dynamics. A modified Bellman-based operator is introduced to support reinforcement learning (RL) policy updates while leveraging prior trajectory information for improved convergence. GMP3 is designed as a distributed framework in which agents influence each other and share policy information along the trajectory: each agent refines its assigned segment and shares with its neighbors via a consensus-based scheme, enabling cooperative policy updates and convergence toward a path shaped globally even under kinematic constraints. We also propose DroneManager, a modular ground control software that interfaces the planner with real UAV platforms via the MAVLink protocol, supporting real-time deployment and feedback. Simulation studies and indoor flight experiments validate the effectiveness of the proposed method in constrained 3D environments, demonstrating reliable obstacle avoidance and smooth, feasible trajectories across both position and orientation. The open-source implementation is available at https://github.com/Domattee/DroneManager.}, language = {en} } @article{KreilingerZottHemmertetal.2026, author = {Kreilinger, Laurens and Zott, Stefan and Hemmert, Werner and Karg, Sonja}, title = {Dry electrode impedance: a new approach towards improved characterization}, volume = {12}, pages = {025005}, journal = {Biomedical Physics \& Engineering Express}, number = {2}, publisher = {IOP Publishing}, address = {Bristol}, issn = {2057-1976}, doi = {https://doi.org/10.1088/2057-1976/ae3e9c}, year = {2026}, abstract = {Electrode-skin impedance plays a crucial role in electrophysiological signal acquisition, influencing signal quality and measurement reliability. We designed a reproducibility measurement setup, using a membrane with a saline solution and a three-electrode Electrochemical Impedance Spectroscopy measurement setup (range 1 Hz-20 kHz), to mimic the electrode-skin impedance. The system allowed controlled application of pressure to the working electrode (WE) and measurement of all setup parameters. With this setup, reproducible results were achieved, with a standard deviation of 5.5\% of the mean impedance across three builds. Potentiostatic and impedance analyzer measurements with six types of dry electrodes produced comparable results, with an average error of 10\%. The six dry electrode types exhibited impedance variations of up to a factor of 10,000 at low frequencies, depending on material and geometry. Ag/AgCl-coated electrodes exhibited an impedance reduction by a factor of 100 at 1 Hz compared to their uncoated counterparts. The proposed setup provides a standardized and reproducible approach for characterizing electrode impedance across different materials, coatings, and geometries.}, language = {en} } @inproceedings{JimenezReyesdelosMozosPereaPanosetal.2025, author = {Jimenez, Victor and Reyes de los Mozos, Mario and Perea Pa{\~n}os, Pau and Fritzsche, Paula Cecilia and Mayer, Kevin and Volkersdorfer, Tina and Hof, Hans-Joachim and Couturier, Christophe and Ernst, Thierry and Zhang, Miao and Saied Mohamed, Mohamed and Rodr{\´i}guez-Arozamena, Mario and Aranguren-Mendieta, I{\~n}igo and P{\´e}rez, Joshu{\´e} and Jousse, Adrien and Murguia, Carlos and van de Wouw, Nathan and Bellessort, Romain and Salami, Behzad and Jevtić, Aleksandar and Bannour, Boutheina and Rodr{\´i}guez Recasens, Manel and Ropero, Isaac and Ozbay, Burcu and Eren, Ali and Bektas, Mustafa and Tezcan, Deryanur and Pilz, Christoph and Haas, Sarah and Lenz, Gernot}, title = {SELFY - Self Assessment, Protection and Healing Tools for a Trustworthy and Resilient CCAM}, booktitle = {Transport Transitions: Advancing Sustainable and Inclusive Mobility, Proceedings of the 10th TRA Conference, 2024 Dublin, Ireland - Volume 6: Connected Mobility Ecosy}, editor = {McNally, Ciaran and Carroll, P{\´a}raic and Martinez-Pastor, Beatriz and Ghosh, Bidisha and Efthymiou, Marina and Valantasis-Kanellos, Nikolaos}, publisher = {Springer}, address = {Cham}, isbn = {978-3-032-06763-0}, doi = {https://doi.org/10.1007/978-3-032-06763-0_106}, pages = {745 -- 751}, year = {2025}, abstract = {SELFY envisions an agnostic toolbox for the self-management of security and resilience of the CCAM (Connected, Cooperative and Automated Mobility) ecosystem, which can be easily deployed to extend the current Operational Design Domain (ODD), providing self-awareness, self-resilience and self-healing mechanisms and enhancing trust between stakeholders. SELFY is based on four pillars: Situational awareness, Resilience, Secure Data Sharing and Trust and provides three groups of tools. SACP (Situational Awareness and Collaborative Perception) tools aim at providing all CCAM actors with a comprehensive understanding of their environment, i.e., the perception of objects, such as other traffic participants and stationary objects. CRHS (Cooperative Resilience and Healing System) tools enable self-protection actions whenever a compromising situation is detected in relation to assets, vehicles, operations, or the system itself. TDMS (Trust and Data Management System) tools establish a secure and trusted environment for data in a collaborative and cooperative context, both for infrastructure and assets, as well as for citizen's data, such as drivers or pedestrians with special attention to privacy considerations. By defining a collaborative environment between the different tools to respond to new threats, risks and attacks SELFY facilitates the comprehension of new challenges in the cybersecurity aspect of CCAMs.}, language = {en} } @inproceedings{JimenezReyesdelosMozosPereaPanosetal.2024, author = {Jimenez, Victor and Reyes de los Mozos, Mario and Perea Pa{\~n}os, Pau and Fritzsche, Paula Cecilia and Mayer, Kevin and Volkersdorfer, Tina and Hof, Hans-Joachim and Couturier, Christophe and Ernst, Thierry and Zhang, Miao and Saied Mohamed, Mohamed and Rodr{\´i}guez-Arozamena, Mario and Aranguren-Mendieta, I{\~n}igo and P{\´e}rez, Joshu{\´e} and Jousse, Adrien and Murguia, Carlos and van de Wouw, Nathan and Bellessort, Romain and Salami, Behzad and Jevtić, Aleksandar and Bannour, Boutheina and Rodr{\´i}guez Recasens, Manel and Ropero, Isaac and Ozbay, Burcu and Eren, Ali and Bektas, Mustafa and Tezcan, Deryanur and Pilz, Christoph and Haas, Sarah and Lenz, Gernot}, title = {SELFY - Self Assessment, Protection \& Healing Tools for a Trustworthy and Resilient CCAM}, booktitle = {Proceedings of the Transport Research Arena (TRA2024) Conference}, publisher = {Zenodo}, address = {Genf}, doi = {https://doi.org/10.5281/zenodo.10978183}, year = {2024}, abstract = {SELFY envisions an agnostic toolbox for the self-management of se-curity and resilience of the CCAM (Connected, Cooperative and Automated Mo-bility) ecosystem, which can be easily deployed to extend the current Operational Design Domain (ODD), providing self-awareness, self-resilience and self-heal-ing mechanisms and enhancing trust between stakeholders. SELFY is based on four pillars: Situational awareness, Resilience, Secure Data Sharing and Trust and provides three groups of tools. SACP (Situational Awareness and Collabo-rative Perception) tools aim at providing all CCAM actors with a comprehensive understanding of their environment, i.e., the perception of objects, such as other traffic participants and stationary objects. CRHS (Cooperative Resilience and Healing System) tools enable self-protection actions whenever a compromising situation is detected in relation to assets, vehicles, operations, or the system itself. TDMS (Trust and Data Management System) tools establish a secure and trusted environment for data in a collaborative and cooperative context, both for infra-structure and assets, as well as for citizen's data, such as drivers or pedestrians with special attention to privacy considerations. By defining a collaborative en-vironment between the different tools to respond to new threats, risks and attacks SELFY facilitates the comprehension of new challenges in the cybersecurity as-pect of CCAMs.}, language = {en} } @article{GleissnerKotonski2025, author = {Gleißner, Christian and Kotonski, Julia}, title = {The classification of rigid torus quotients with canonical singularities in dimension three}, volume = {36}, journal = {Rendiconti Lincei: Matematica e Applicazioni}, number = {2}, publisher = {EMS Press}, address = {Berlin}, issn = {1720-0768}, doi = {https://doi.org/10.4171/RLM/1070}, pages = {315 -- 352}, year = {2025}, abstract = {We provide a fine classification of rigid 3-dimensional torus quotients with isolated canonical singularities, up to biholomorphism and diffeomorphism. This complements the classification of Calabi-Yau 3-folds of type III0​, which are those quotients with Gorenstein singularities.}, language = {en} } @inproceedings{RadtkeHuberBock2023, author = {Radtke, Maximilian-Peter and Huber, Marco F. and Bock, J{\"u}rgen}, title = {Increasing Robustness of Data-Driven Fault Diagnostics with Knowledge Graphs}, booktitle = {Proceedings of the Annual Conferenceof the PHM Society 2023}, editor = {Kulkarni, Chetan S. and Roychoudhury, Indranil}, publisher = {PHM Society}, address = {Rochester}, doi = {https://doi.org/10.36001/phmconf.2023.v15i1.3552}, year = {2023}, abstract = {In the realm of prognostics and health management (PHM), it is common to possess not only process data but also domain knowledge, which, if integrated into data-driven algorithms, can aid in solving specific tasks. This paper explores the integration of knowledge graphs (KGs) into deep learning models to develop a more resilient approach capable of handling domain shifts, such as variations in machine operating conditions. We present and assess a KG-enhanced deep learning approach in a representative PHM use case, demonstrating its effectiveness by incorporating domain-invariant knowledge through the KG. Furthermore, we provide guidance for constructing a comprehensive hierarchical KG representation that preserves semantic information while facilitating numerical representation. The experimental results showcase the improved performance and domain shift robustness of the KG-enhanced approach in fault diagnostics.}, language = {en} } @article{DaSilvaJuniorMuellerBirkneretal.2026, author = {Da Silva Junior, Amauri and M{\"u}ller, Steffen and Birkner, Christian and Jazar, Reza and Marzbani, Hormoz}, title = {Fault Tolerant Control With Reinforcement Learning for Evasive Maneuvers Using a Scaled Vehicle}, volume = {14}, journal = {IEEE Access}, publisher = {IEEE}, address = {New York}, issn = {2169-3536}, doi = {https://doi.org/10.1109/ACCESS.2026.3661179}, pages = {21353 -- 21383}, year = {2026}, abstract = {Autonomous vehicle controllers are responsible for handling the vehicle at all times in any situation, including emergency conditions. An emergency might arise from, e.g., adverse weather conditions, short-time detection, and system faults. In this paper, we develop a fault-tolerant controller to handle actuator faults for an over-actuated autonomous vehicle based on reinforcement learning. A worst-case scenario is selected for the controller development, involving short-time detection of the preceding objects, high velocity, and dry to wet road conditions. The design of the vehicle controller is performed in three steps. First, a robust controller based on sliding mode control with lateral and longitudinal coupled strategy was built to ensure stability in emergency scenarios. Secondly, a strategy was proposed to identify the most critical vehicle faults that might lead to a crash. Building on these foundations, this study extends the vehicle controller to handle vehicle faults with a reinforcement learning strategy, enabling adaptive and robust fault handling in complex fault scenarios. The vehicle controller is designed and optimized in IPG-Carmaker®, and proof of concept is carried out in a scaled 1:3.33 test vehicle. The results demonstrate the robustness of the proposed controller in an emergency single-lane change with a velocity of up to 130 km/h. Tests in the scaled vehicle demonstrate the vehicle controller's accuracy against simulation, with the application of reinforcement learning strategy in real-case scenarios.}, language = {en} } @article{VogtMuellerKampaetal.2021, author = {Vogt, Anna and Mueller, Ralph and Kampa, Thomas and Stark, Rainer and Großmann, Daniel}, title = {Concept and architecture for information exchange between digital twins of the product (CPS) and the production system (CPPS)}, journal = {Procedia CIRP}, publisher = {Elsevier}, address = {Amsterdam}, issn = {2212-8271}, doi = {https://doi.org/10.1016/j.procir.2021.11.217}, pages = {1292 -- 1297}, year = {2021}, abstract = {The Digital Twin concept and CPS- and IIoT-based approaches are increasingly important topics concerning future Industry 4.0 architectures. They offer high potential for dynamical aspects in intelligent production planning and control as well as part traceability and documentation. Standardized information exchange is an upcoming requirement among the whole supply chain. This paper presents a concept for a Digital Twin architecture based on motor production in the automotive industry. The key aspect is an information exchange structure for Digital Twins of products and production systems that are combined using principles of Dynamic Aggregation.}, language = {en} } @unpublished{DoenmezTejerodelaPiedraKloseetal.2026, author = {D{\"o}nmez, {\"O}mer and Tejero de la Piedra, Ricardo and Klose, Simona and Riolet, Matthieu and Rozek, Lukas and Vaculin, Ondrej and Hach, Christian}, title = {Approach for Passive Safety Assessment of Rearward-Sitting Occupants}, publisher = {Zenodo}, address = {Genf}, doi = {https://doi.org/10.5281/zenodo.18338626}, year = {2026}, abstract = {The introduction of highly automated vehicles (HAVs) will allow vehicle occupants to take advantage of new seating configurations, such as sitting rearward in the first row. One critical aspect of assessing occupant safety during high-speed impacts is the lack of a dedicated safety framework for rearward-facing passengers in the first row. This paper introduces a method to develop new assessment criteria for these novel seat configurations. Thus, this research presents some preliminary results of rearward-facing occupant injury biomechanics analyses carried out employing a variety of anthropomorphic test devices (ATDs) and the VIVA+ 50M human body model (HBM), restrained with different belt configurations and considering different seat typologies. It reviews the suitability of 50th percentile male ATDs to capture a biofidelic engagement with the seat structure and belt system and evaluates the reaction loads on the occupant, along with the energy management resulting from seat back rotational stiffness and energy-absorbing foams layered behind the seat cushion. Based on the results, the THOR-AV-50M is a suitable candidate for further biofidelity analysis. Torso occupant loads can be effectively reduced utilizing seat back rotation but pelvis load management requires further studies.}, language = {en} }