Modeling mental spatial knowledge processing An AI perspective

Abstract

This chapter addresses mental spatial knowledge processing from an artificial intelligence (AI) point of view. It first reviews the characteristics of mental representation structures used for mental reasoning processes, motivates mental knowledge processing as a construction process, and points to the use of external media as an important factor in dealing with more complex spatial problems. A range of models of intelligent spatial information processing has been proposed both in psychology and in AI. After giving an overview of selected models, the architecture Casimir is presented as a framework for modeling spatial reasoning with mental models and mental images. On the basis of this architecture, the design of representation structures in working memory, the task of computationally modeling mental processing of shape information, and the issue of controlling mental resources in reasoning processes are discussed as challenging issues from an AI perspective. The chapter closes with some considerations regarding the assessment and validation of computational models of mental spatial knowledge processing.