Guest lecture: Utilising domain knowledge and spatial reasoning: the CLP(QS) spatial reasoning system and its application in architectural design

Dr. Carl Schultz, Institute for Geoinformatics, University of Münster, Germany.

13.01.2017 | Jens Kargaard Madsen

Dato man 16 jan
Tid 13:00 14:00
Sted Room 430E, Finlandsgade 22, 8200 Aarhus N

Abstract:
Researchers in the field of Knowledge Representation and Reasoning (within Artificial Intelligence) aim to formally capture the way that humans conceptualise and reason about objects and events in specific application domains; this enables software systems to get much greater leverage from input data (e.g. images, videos) by exploiting background domain knowledge. A critical subfield is spatial reasoning that enables software systems to check whether a model of a scene is semantically consistent, perform more robust scene interpretation, and provide higher-level interfaces for querying such models of a scene. While spatial reasoning is a promising field, there are a number of critical barriers that have, thus far, prevented its wide deployment in industry and other areas of AI.

In this presentation I will introduce the field of declarative spatial reasoning, and present an overview of my research in this area. I am developing a spatial reasoning system CLP(QS) that directly addresses two critical issues in the field, namely (a) combining geometric information and qualitative spatial information, and (b) integrating spatial reasoners in larger software systems. I will then present results on applying spatial reasoning in architectural design within the DesignSpace project, where I am developing building design analysis and decision support tools for architects (based on formal theories of visibility, movement, affordance, lighting, and so on) in collaboration with industry partners at the New Parkland Hospital in Dallas, currently the largest funded hospital construction project in the US.

Carl Schultz is a lecturer and researcher at the Institute for Geoinformatics at the University of Münster, Germany. Previously he was a Postdoctoral researcher in the Cognitive Systems Group (CoSy) at the University of Bremen, Germany. In 2010 he completed his Ph.D. at the University of Auckland, New Zealand, on software engineering methodologies for developing spatial reasoning applications. His research aims to develop methods for geometric and qualitative spatial reasoning in the context of Artificial Intelligence and Knowledge Representation and Reasoning, with a particular focus on software engineering methodologies and real-world, industry-scale applications. 

Forelæsning / foredrag