Scientific Networking Days 2025: Revisiting questions, refining methods, and building connections

Poster sessions as a space for in-depth dialogue

One of the central features of the Networking Days was again the internal poster session. With 17 posters presented by cluster members, the sessions created a focused yet open setting where individual projects could be discussed in more detail. Unlike oral talks, the poster format allowed for direct conversations in small groups, offering time to dig deeper into data, design choices, and conceptual frameworks. Researchers used the sessions to ask questions, challenge assumptions, and test the clarity of their own thinking in conversation with others.

The presentation competition highlighted three particularly resonant contributions. The first prize went to Félicie Dhellemmes and Valerii Chirkov, who presented a study linking behavioral fieldwork on group search in natural environments with modeling approaches in artificial agents. Their work offered insights into how search strategies scale from individuals to collectives, and how such dynamics might inform future designs in robotics and behavioral modeling.

Second place was awarded to Furkan Davulcu and Bassel Katamish for their presentation on how intelligent agents can generate robust real-world behavior by exploiting regularities at multiple levels of abstraction. They propose that the ability to act intelligently depends on identifying and using these regularities in the world. Their approach frames hierarchical abstraction as a fundamental principle of intelligence, providing a foundation for multi-layered decision-making in both biological and artificial systems.

Anna Lange and Helene Ackermann received third place for their project investigating adaptive behavior in human-robot interaction. Their work combined cognitive modeling with experimental methods to explore how social robots can adjust to individual learners and support learning through more personalized interaction.

Reviewing and realigning: the general assembly

Alongside research presentations and poster discussions, the general assembly provided a space to take stock of broader developments within the cluster. Updates from the Executive Board and the Coordination Office offered transparency on governance and upcoming milestones. Committees reported on their ongoing work, ranging from curriculum and ethics to internal communication and support structures. The assembly created a shared picture of the cluster’s achievements over the past years.

Marking a milestone: the successful completion of doctoral programs

This year’s Networking Days included a moment of recognition for eight researchers who completed their interdisciplinary doctoral qualification program within the cluster. Their projects reflect the diversity and depth of the research environment that has taken shape over the past years.

Manuel Baum studied how robots and animals gather information through interaction when faced with unfamiliar mechanical problems. His research focused on task-directed exploration and problem solving, combining robotics with behavioral biology.

Florian Blume worked on face perception, using deep neural networks to explore how prior knowledge influences visual processing. His work connected machine learning with cognitive models of recognition.

Danny Driess focused on integrating perception and planning in robotic manipulation. His work advanced methods in task and motion planning by learning how to couple sequential decision-making with responsive execution.

Fritz Francisco used fish schools to study social learning and behavioral adaptation in dynamic environments. His findings offered insights into how animals transfer information and reduce uncertainty—knowledge that can also inform artificial systems.

Mohsen Raoufi worked on differences in robotic collective estimation by investigating how differences within collectives, ranging from inter-individual variations and functional roles to informational diversity and heterogeneous network structures, shape the dynamics of collectives, in particular for opinion dynamics and estimation scenarios.

Nicolas Roth investigated how humans visually search and interpret complex scenes. His work combined computational modeling, neuroscience, and visual cognition to improve both understanding and application.

Lynn Schmittwilken explored how the human visual system perceives three-dimensional structure. Her research aimed to better understand depth processing and how these mechanisms might inform computer vision.

Hae Seon Yun contributed to SCIoI’s conceptual and technical work on learning companions, focusing on how social and emotional factors can support human learning. She also played a central role in shaping the integration project.

Vincent Wall worked on soft robotics, developing new versions of the RBO Hand and advancing sensor technologies for adaptive robotic manipulation.

Keeping the conversation going

As an annual scientific event, the Scientific Networking Days are about process, about checking in, identifying gaps, sharpening tools, and refining questions. They are about seeing how one project might speak to another, and how individual work contributes to a shared effort. This year’s event once again showed how science can thrive in a place where interdisciplinarity is a working practice.