Action selection based on multi-modal anchoring

The project aims at developing a computational theory of the formation and multi-modal perceptual anchoring of symbolic concepts. This will be explored in a scenario in which a rat is building up a multimodal representation of a room with object-like landmarks while performing a task. The principles extracted from these experiments will be generalised step-by-step to more communicative situations with human children, adolescents, and younger and older adults. In parallel, it will be implemented and evaluated on a robotic platform. Key questions of the project are:

• What are the bottom-up and top-down mechanisms for the formation and anchoring of symbolic concepts in animals and humans
• How do these mechanisms change in during the cognitive development from childhood through adulthood to older age?
• Does a fusion of different sensory modalities take place on a sub-symbolic or symbolic level?

These questions will be studied from different inter-disciplinary perspectives:

• Behavioural studies with rats will provide insights how far multi-modal percepts will be differentiated into the single modalities and if the joint perception of coherent percepts will lead to a significant boost of attraction/distraction.
• The experiments in humans will take a further step towards the ontology of semantics as we expect insights into children and adults’ pre-linguistic meaning processing.
• Based on these behavioural results from rats and humans, this project will develop a computational model for the multi-modal anchoring of both known and novel symbols in robotic systems.

[view:groupmembers==191]