Architecture of attentional processes in active vision


This project bridges the analytic disciplines psychology and neuroscience with synthetic approaches (i.e., computational models).

We aim at a better understanding of selective processes that promote efficient resource allocation for perception, memory, and motor control in active vision.

PI: Martin Rolfs

Collective intelligence and decision making with application to medical information processing


This project bridges the areas of psychology, biology and has immediate implications for all domains in which individuals and collectives of individuals make consequential decisions such as in medical diagnostics.

The key question is under what circumstances does it pay to combine the intelligence of individuals because the collective can outperform the best individual.

PIs: Ralph Hertwig, Jens Krause, Thorsten Pachur

Communication: interplay between verbal, visual and social-emotional influences

right-30 Image copyright: © SEPS:Curtis Publishing, Indianapolis

Human communication involves the simultaneous processing of information from different modalities and the coordinated use of core verbal and non-verbal cognitive functions, including language production and comprehension, face identity and expression perception, semantic and emotion processing, and social cognition. The project aims at understanding and describing these core functions and their interactions from a social-communicative perspective.

PI: Rasha Abdel Rahman(Image copyright: © SEPS:Curtis Publishing, Indianapolis)

Deep learning with robotic priors; combining deep learning and algorithms

right-30 Image copyright: CC-BY-SA

This project lies at the intersection of robotics and machine learning. The goal is to enable robots to learn in a data-efficient way by providing certain prior knowledge. More specifically, we provide the robot with algorithms and let it learn to use those using deep learning.

PI: Oliver Brock(Image copyright: CC-BY-SA)

Effects of social learning environments on cognitive development


This field of research investigates the role that social learning environments (e.g., instructional settings) play for developmental dynamics in human cognition. It bridges the areas of psychology, educational sciences and robotics by developing mathematical models of processes that define the transfer between teachers and learners.

PI: Rebecca Lazarides

Influence of emotions on information processing


Integrating complex environmental and social information into decision making can be facilitated by reducing dimensionality through emotions. By analysing and experimentally testing living as well as artificial model organisms we will explore possibilities of cost-reduction in artificial cognition (e.g., battery, memory, processor and sensory capabilities).

PIs: Jens Krause, Lars Lewejohann, Pawel Romanczuk, Christa Thöne-Reineke

Influences of prior knowledge on face and object perception


The project combines psychological and neuroscientific approaches to understand how effects of different domains of knowledge (e.g., linguistic, functional-semantic, social and affective knowledge) on low and high-level stages of face and object perception.

PI: Rasha Abdel Rahman

Investigating the Dynamics of Self and Social Cognition


This project combines philosophical inquiries on self and social cognition with research in developmental robotics. Drawing on research from philosophy and human psychology, we aim at a better understanding of the dynamics and processes of group cognition and collaboration in synthetic multi-agent scenarios.

PIs: Verena Hafner, Miriam Kyselo, Michael Pauen

Iterative learning control to adapt neuroprostheses to individual users


This project uses Iterative Learning Control (ILC) to develop adaptive neuroprostheses that can be used in stroke rehabilitation. It aims at automatically adjusting the intensity of artificial muscle stimulation to induce near physiological repetitive movements. Standard ILC, however, requires rather strict assumptions as, e,g., identical duration of all cycles in periodic motions. As this is rarely true in biomedical applications, an integral part of this project has been to appropriately extend the underlying theory.

PI: Jörg Raisch

Life transitions and adaptive development of socio-cognitive characteristics


The project connects theoretical approaches from educational science, psychology, sociology and robotics by studying how the interaction between a learner and a teacher during life transitions affect the development of socio-cognitive characteristics.

PI: Rebecca Lazarides

Manual dexterity in humans and robotic systems and associated cognitive abilities


Human grasping skills are far superior to those of robots. We investigate the principles that lead to this superior performance in human grasping and work towards transferring these principles to robotic systems. Our key hypothesis is that human grasping performance crucially depens on the purposeful exploitation of contact with the environment.

PIs: Oliver Brock, Marianne Maertens

Models of collective cognition in swarms


The project connects the synthetic disciplines of computer science, applied mathematics and theoretical physics with behavioral biology. This is done by exploring collective decision making and information processing in animal groups, developing mathematical models, and extracting biologically inspired algorithms of collective cognition.

PIs: Jens Krause, Pawel Romanczuk

Physical Exploration Challenge


One of the hallmarks of human development is exploratory behavior. In this project, we develop robots that explore the degrees of freedom in their environment (door handles, drawers, scissors, etc.). To generate such behavior we have to address questions of representation, action selection, and embodiment.

PI: Oliver Brock

Shape priors for 3D object recognition


3D object recognition requires comparison of currently perceived shape with a priori information. In ongoing research, we learn shape priors from training objects showing that the estimated shape prior is capable to express fine details to a certain degree. By applying shape priors to technical measurements the accuracy and completeness of reconstructed models can be significantly increased. Future work will investigate how humans make use of shape priors when perceiving shape, and how technical representations differ from biological ones.

PI: Olaf Hellwich