Adoption of Deep Generative Models in Architectural and Urban-Form Generation

Integrating deep generative models (DGMs) into architectural and urban-form generation is an innovative approach to support the design process. However, whether these applications are what designers need is rarely considered. To capture the user acceptance and visions of applying DGMs to architectural and urban-form generation, survey research was conducted. Drawing on the technology acceptance model (TAM), this study examines factors affecting user acceptance. TAM in this application technology is validated. The results indicate participants' evaluation of perceived usefulness (PU), attitude towards use (ATU), and intention to use (ITU) are modestly positive, while that of perceived ease of use (PEOU) is neutral, in terms of adopting DGMs in architectural and urban-form generation. Participants further display a mild and positive vision of integrating topology, space syntax, and typology in future adoption. Based on the result, a conceptual framework of DGMs-aided architectural and urban-form generation is proposed. The framework can be developed into different versions for beginners, intermediates, and experts in distinct areas considering their local cultural backgrounds to improve user-friendliness. The findings provide actionable insights for the development of user-focused DGM tools in architectural and urban-form generation.