In the EU FP7 project CogX our group has been in charge of cross-modal learning and the development of an interactive system for learning in dialogue with a tutor. In this paper we present representations and mechanisms that facilitate continuous learning of visual concepts in dialogue with a tutor and show the implemented robot system. We present how beliefs about the world are created by processing visual and linguistic information and show how they are used for planning system behaviour with the aim at satisfying its internal drive -- to extend its knowledge. The system facilitates different kinds of learning initiated by the human tutor or by the system itself. We demonstrate these principles in the case of learning about object colours and basic shapes. This distributed, heterogeneous, and very complex system has been developed in collaboration with other project partners under our guidance and supervision, and will serve as a basis for further research and development of artificial cognitive systems in our group.
F.24 Improvements to existing system-wide, normative and programme solutions, and methods
COBISS.SI-ID: 8881748In this talk we present our approach to visual categorization which combines a learned compositional hierarchy, representing shapes of multiple object classes, and a coarsetofine matching scheme that exploits a taxonomy of objects to perform efficient object detection. From simple contour fragments frequent spatial configurations are learnt and recursively combined into increasingly more complex and classspecific shape compositions layer by layer. To speed up recognition, a taxonomy of constellation models cascaded from coarsetofine resolution is constructed automatically in a way that inimizes the number of expected computations during recognition by optimizing the costtopower ratio.
B.04 Guest lecture
COBISS.SI-ID: 8991316Dr. Bojan Kverh is employed in commercial software development.
F.17 Transfer of existing technologies, know-how, methods and procedures into practice