The What, Why, What-If And How-To For Designing Architecture, Explainability For Auto-Completion Of Computer-Aided Architectural Design Of Floor Plan Layouting During The Early Design Stages

Jessica Bielski Technical University of Munich
Christoph Langenhan Technical University of Munich
Christoph Ziegler Technical University of Munich
Viktor Eisenstadt DFKI/ University of Hildesheim
Frank Petzold Technical University of Munich
Andreas Dengel DFKI/ University of Hildesheim
Klaus-Dieter Althoff DFKI/ University of Hildesheim

In the next thirty years, the world’s population is expected to increase to ten billion people, posing major challenges for the construction industry. To meet the growing demands for residential housing in the future, architects need to work faster, more efficiently, and more sustainably, while increasing architectural quality. The hypothetical intelligent design assistant WHITE BRIDGE, based on the methods of the ‘metis’ projects, suggests further design steps to support the architectural design decision-making processes of the early design phases. This facilitates faster and better decisions early in the process for a more responsible resource consumption, better mental well-being, and ultimately economic growth. Through a case study we investigate if additional information supports the understanding of these suggestions to reduce the cognitive workload of architectural design decisions on the backdrop of their respective representation. The paper contributes an approach for visualising explanations of an intelligent design assistant, their integration into paper prototypes for case studies, and a workflow for data collection and analysis. The results suggest that the cognitive horizon of the architects is broadened by the explanations, while the visualisation methods significantly influence the usefulness and use of the conveyed information within the explanations.

Keywords: Explainability, Artificial Intelligence, Xai, Sdg 3, Sdg 8, Sdg 12

View Full Paper