Workshop #1 – Towards Closer German French Cooperation
Dr. Pierre Alliez (Inria), Prof. Dr. Philipp Slusallek (DFKI) / Location: Room "Reuse", DFKI
Germany and France are moving closer together in the field of artificial intelligence. But what are our respective strengths? Which AI topics unite us? What opportunities do the AI gigafactories unlock? And what tangible benefits value does our cooperation create - for the economic sectors and public authorities of both countries as well as for Europe as a whole? The workshop brings together participants from the French AI clusters and the German AI competence centres. The aim is to generate enthusiasm for cross-border research and to launch joint project ideas that have the potential to strengthen European innovative capacity in the long term.

Workshop #2 – Forum on Promoting, Training, and Managing Doctoral Candidates
Filippo De Bortoli and Dr. Frank Loebe (ScaDS.AI Dresden/Leipzig), Dr. Tina Schwabe (BIFOLD) / Location: IWi Meeting Room 0.08, DFKI
Doctoral Candidates (DCs) represent a large part of the members of our AI centres and significantly contribute to their output. Are we giving them the right support? In this forum, we seek exchange with all those involved in coordination and support of DCs as well as DCs and supervisors, on topics such as working conditions, training formats and provision, supervision agreements, management tooling and processes. Further topics are welcome! Participation is open to all attendees of the AHM.

Workshop #3 – Certificate course „Responsible AI”
Prof. Birte Platow (ScaDS.AI Dresden/Leipzig, TU Dresden) / Location: Room "Bledsoe", DFKI
Each discipline has its own ideas about what “responsible AI” means—privacy, security, transparency, explainability, regulatory compliance, a specific mindset, and so on. In the planned certificate course “Responsible AI,” we want to make these different approaches accessible to a wider audience AND, at the same time, use the idea of research-based teaching and learning to gain new insights for interdisciplinary research into a holistic understanding of responsible AI. In the workshop, we will present the concept of the course and develop it further with you.

Workshop #4 – Multi-Modal Biosignal Disentanglement
Dr. Theekshana Dissanayake and Alexander von Lühmann (BIFOLD) / Location: Room "Turing" I+II, DFKI
Disentangled representation learning of biosignals is a central challenge in computational neuro-science and the development of intelligent wearable technologies. Whether the goal is to separate brain activity from systemic physiological confounders in neurological data or to isolate subject-specific patterns in biosignals for improved generalizability, effective disentanglement is essential for advancing our understanding of complex neuro-physiological interactions within the human body. This workshop will bring together researchers exploring methodological strategies for biosignal disentanglement, including—but not limited to—representation learning, generative modeling, and graph-based approaches. Please reach out to thilakasiri.dissanayakage@tu-berlin.de to coordinate your contribution. Based on the contributions, we want to use the remaining time for a collaborative discussion, knowledge exchange, and networking across AI centres.

Workshop #5 – AI for the Benefit of Society
Dr. Clemens Ackerl (SERI), Alexander Wilhelm (TrADeMaS), Dr. Solvejg Gunkel (GenKI-IR), Prof. Matthias Uhl (Ethyde), Dr. Ulrike Völlinger (DLR-PT) / Location: VisRoom, DFKI
In interdisciplinary projects, the BMFTR funds the investigation of ethical, legal, sociological, political and economic aspects raised by the use of AI systems. The projects accompany the technical AI-research projects. In the workshop, four projects will present their research results for discussion. The workshop serves to network the research projects with each other and to discuss new research questions. It is open to all interested parties.

Workshop #6 – Amortized and In-Context Simulation-Based Inference for Dynamical Systems: From Principles to Scientific Discovery
Dr. Ramsés Sánchez and Patrick Seifner (LAMARR) / Location: Room "Reuse", DFKI
In recent years, Bayesian inference has witnessed a significant shift: The emergence of neural network models pretrained on synthetic data to enable in-context (or zero-shot) inference. Prior-Fitted Networks (PFNs), for instance, approximate posterior predictive distributions in a single forward pass. Foundation Inference Models (FIMs) similarly operate in zero-shot mode but focus instead on estimating the latent functions that govern the evolution of unseen dynamical systems. Both approaches fall under the broader umbrella of Simulation-based Inference (SBI). This workshop explores these emerging paradigms and their role in enabling automated scientific discovery. We aim to examine their two key components: first, the design of synthetic prior training distributions as a vehicle for encoding inductive biases, including the potential use of information-theoretic principles to guide their construction; and second, the architectural and computational challenges involved in building models capable of efficiently processing large context sets.

Workshop #7 – Stronger Together: Building cross-institutional Gender Equality and Diversity movement in the German AI centers
Dr. Alexandra Stang (MCML), Dr. Laura Wollenweber (BIFOLD) / Location: Room "Turing" I+II, DFKI
This workshop brings together researchers and staff to strengthen gender equality and diversity across all AI competence centers. After an inspiring talk from an expert in women’s networks, Maria Petrova-El Sayed, participants are invited to collaborate in 3 table discussions: Networks that Empower – Building a cross-center women’s network; Hiring for Diversity – Designing diversity-sensitive recruitment; Inclusive by Design – Fostering diversity-sensitive teams.

Workshop #8 – Discussion of Joint Transfer Activities of the German AI Research Centres
Dr. Jack Thoms (BIFOLD) / Location: VisRoom, DFKI
We want to propose and discuss possible joint transfer activities among the AI research centres. Specifically we want to propose:
- A joint initiative/campaign together with the FUTURIUM and the Weizenbaum Institute on “Future of Men-Machine-Relationship” for 2026
- Joint efforts in the initiative “KI macht Schule”
- Aligning efforts in national AI startup initiatives such as the Startup Factories and AI Nation.
Other proposals for joint transfer activities are very welcome!