XVIII. MİMARLIKTA SAYISAL TASARIM ULUSAL SEMPOZYUMU
An Evaluation on Machine Learning and New Approaches in the Production of Architectural Representation
Yayıncı:
Balıkesir Üniversitesi Yayınları
With the development of computational design theory, rule-based systems began to be used in the field of design and a new environment for creating architectural representation in which human-computer interaction is at the forefront was born. In this period when the human-centered design approach has transformed into a design approach managed by the computer, which is a machine at the most basic level, artificial intelligence technology has rapidly changed this mutual production process. It is thought that both disciplines can be seen as technologies that can be used to define and produce architectural representation and make the design process original and practical. Especially in the last 10 years, machine learning, a product of artificial intelligence, has been managing this process in which new design approaches are developed. Thanks to the discovery of artificial intelligence-supported productive design tools, many architects have, at least partially, stayed out of construction problems and taken part in a relatively autonomous architectural practice that develops in the digital environment, and the concept of design, design tools and designer identity have begun to transform. Today, the design process in architecture and the production methods of architectural representation are changing with increasing momentum, and human-made creative design methods are being replaced by digital design tools produced by artificial intelligence. An increasingly accelerating design practice has developed, focusing on the process rather than the final product, facilitating the process in this respect, but on the other hand, a critical design practice has developed in terms of many parameters related to the final product, such as originality, originality, innovation and repeatability. In this study, it is aimed to reveal the change in the architectural design process and the production method of architectural representation, to draw attention to the advantages it can provide in the field of architecture and the potential for its use, and to create an environment for discussion on the subject within the framework of the professional discipline. It is intended to bring to the agenda new approaches that use machine learning tools in the production of architectural representation and their important effects on architectural practice. In this regard, a bibliometric analysis was carried out on research on the radical and rapid transformation of the design process and the production tools of architectural representation in the field of architecture. In the light of the research examined, the change of the design concept, design process and design product was interpreted. Thus, it is aimed to draw attention to the interest in the field in the architectural literature, to question different ways of addressing it, and to provide the opportunity to think about attitudes towards the adaptation and effectiveness of artificial intelligence-oriented productive design tools into architectural practice. Examining the concept of digital design within the framework of architectural theory, conceptual change of design parameters with data-based design, and attitudes towards new design approaches in the architectural literature were evaluated. It is thought that such a perspective can theorize this rapid transformation in practice.