Uploaded March 2021 | Updated November 2026, 3 hours ago
Programming an artificial system to function in different situations is challenging. However, this can be facilitated, if we provide it with the ability to perceive and learn the way we do.
Humans, especially infants, learn extensively through exploring causes and effects. They learn what to expect next, and these predictions influence how they further perceive and adapt to new situations.
In the Predictive Robots project, we develop curious artificial systems that learn like infants, building up internal models of their bodily capabilities.
Through these models, robots can learn what to expect. Expected information can be filtered out from incoming sensory data, allowing them to focus their resources on novel or more relevant activities.
Predictive processes thus enable a low-cost enhancement of their perceptual capabilities.
Predictive robots are also more aware of how their activity affects the environment they work in. This advances the research on self-aware artificial systems and, ultimately, helps our understanding of human development.
Programming an artificial system to function in different situations is challenging. However, this can be facilitated, if we provide it with the ability to perceive and learn the way we do.
Humans, especially infants, learn extensively through exploring causes and effects. They learn what to expect next, and these predictions influence how they further perceive and adapt to new situations.
In the Predictive Robots project, we develop curious artificial systems that learn like infants, building up internal models of their bodily capabilities.
Through these models, robots can learn what to expect. Expected information can be filtered out from incoming sensory data, allowing them to focus their resources on novel or more relevant activities.
Predictive processes thus enable a low-cost enhancement of their perceptual capabilities.
Predictive robots are also more aware of how their activity affects the environment they work in. This advances the research on self-aware artificial systems and, ultimately, helps our understanding of human development.










