The course focuses on artificial cognitive systems and in particular on robots, with a special emphasis on the role of natural language within such systems. It is an interdisciplinary course which brings together theories, findings and methodologies from Cognitive Science, Neuroscience, Theoretical and Computational Linguistics and Multimedia Systems.
In using multimodal and multisensory communication examples, the idiosyncratic characteristics and the semantic interaction of natural language with perception and action is presented. This interaction takes place -among others- through “syntax”, a fundamental cognitive mechanism. Syntax is shown to be core not only in verbal but also in sensorimotor behaviour; its relation to long term memory is specifically presented and a semantic memory model for cognitive agents is explained in detail. The latter involves getting familiar with multimodal, referential, and recursive semantic networks. Within this framework, natural language processing is revisited, rendering reference a core requirement in language analysis, that forms a bridge with the sensorimotor aspects of a cognitive system. The course is enriched with examples from two robotic applications: (a) verbal human-robot interaction in everyday life, and (b) visual scene understanding and verbalization.
Indicative Topics:
- Introduction to Artificial Cognitive Systems and Cognitive Robotics
- Cognitive Architectures and the role of Language
- Language, Perception and Action (1): characteristics and relations
- Language, Perception and Action (2): semantic interaction in multisensory and multimodal communication
- Syntax as a Fundamental Cognitive Mechanism: from language grammar to the grammar of action
- Recursion as a Cognitive Phenomenon
- Embodied Language Processing
- Computational Semantic Memory and the Role of Language
- Robotic Applications: (a) Verbal Human-Robot Interaction: from language to action, and (b) Visual Scene Understanding: from image to language
Pastra Katerina
kpastra
athenarc [dot] gr