A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
A Language Prompt Model for Architectural Aesthetics
Our main concern with AI and architecture is that AI will ultimately mimic what we call creativity even though it remains a black box of the type Banham described [1]. We claim to recognize it when we see it but can’t agree what it is, how it works, how to foster it, or how to teach it. If we believe creativity is the generation of novel shapes, then it won’t make any difference if it’s performed wholly or in part by computation, intern farm or some other means of generating quasi-random permutations. Meanings will still need to be assigned to those shapes and it’s the black box that does that.
Any design process produces better results when the instructions are clear. Quickly searching a large image dataset is a computational problem but organizing it and instructing AI to retrieve and synthesize relevant information is an aesthetic one. The author reverse-engineers Charles Jencks’ model of an iconic building into one that also describes (and hence classifies) other aesthetic effects. The names of these effects become language prompts linking the aesthetic problem to the computational one, allowing large image datasets to be consistently indexed and searched, enabling more relevant training, and reducing the need for meanings to be externally assigned.
A Language Prompt Model for Architectural Aesthetics
Our main concern with AI and architecture is that AI will ultimately mimic what we call creativity even though it remains a black box of the type Banham described [1]. We claim to recognize it when we see it but can’t agree what it is, how it works, how to foster it, or how to teach it. If we believe creativity is the generation of novel shapes, then it won’t make any difference if it’s performed wholly or in part by computation, intern farm or some other means of generating quasi-random permutations. Meanings will still need to be assigned to those shapes and it’s the black box that does that.
Any design process produces better results when the instructions are clear. Quickly searching a large image dataset is a computational problem but organizing it and instructing AI to retrieve and synthesize relevant information is an aesthetic one. The author reverse-engineers Charles Jencks’ model of an iconic building into one that also describes (and hence classifies) other aesthetic effects. The names of these effects become language prompts linking the aesthetic problem to the computational one, allowing large image datasets to be consistently indexed and searched, enabling more relevant training, and reducing the need for meanings to be externally assigned.
A Language Prompt Model for Architectural Aesthetics
Lecture Notes in Civil Engineering
Di Marco, Giancarlo (editor) / Lombardi, Davide (editor) / Tedjosaputro, Mia (editor) / McKay, Graham Brenton (author)
xArch – creativity in the age of digital reproduction symposium ; 2023 ; Suzhou, China
2024-02-24
8 pages
Article/Chapter (Book)
Electronic Resource
English
Architectural detailing : function, constructibility, aesthetics
| UB Braunschweig | 2007
Aesthetics of Architectural Structural Detailing
| British Library Online Contents | 1997
Architectural detailing : function, constructibility, aesthetics
| TIBKAT | 2016
Architectural detailing : function constructibility aesthetics
| TIBKAT | 1993
Aesthetics of Architectural Structural Detailing
| Online Contents | 1997