07 October 2019
Free University of Bozen-Bolzano
Multiagent Systems Department
Oliver Kutz

Human common sense is notoriously difficult to capture symbolically and to tackle computationally within AI, and is of particular importance within areas such as cognitive robotics and conceptual modelling.

In this talk, we will introduce two interlinked approaches to the formal modelling of common sense, namely firstly 'Tooth Logic', a new formalism based on description logic that can be used to define concepts declaratively which are learned from weighted features, and secondly ISL, a multi-modal logic for image schemas which captures basic spatio-temporal relations.

Whilst ISL and its agentive extensions can be used to formalise the principles of an agentive naïve proto-physics, tooth logic allows to describe concepts more in line with views from psychology and cognition.

Institution department: 
Research Centre on Knowledge and Data (KRDB)