A carregar...
Publicação
A modelação paramétrica e a geração de imagens por IA como ferramentas de projeto em arquitetura : a computação e a AI como apoio do arquiteto no trabalho de atelier
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 18.82 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A presente Dissertação de Mestrado investiga o potencial da modelação paramétrica e da inteligência artificial (IA) gerativa no âmbito de projeto e design arquitetónico, enfatizando a integração destes processos de trabalho computacional como ferramentas de apoio ao arquiteto em ambiente de atelier. Um trabalho elaborado sob uma componente teórica de investigação de bibliografia e conhecimentos consolidados e uma componente pratica de testagem de processos alternativos de trabalho com base em experiências de trabalho do autor em ambiente de atelier. A componente teórica explora a aplicabilidade e o valor do trabalho projectual sobre um pensamento computacional, que apesar de ser um dos pilares base da Ciência dos Computadores não está diretamente vinculado ao uso do computador. Tendo em conta que Computação é o estudo de processos de “problema-solving”, sendo o acumular multidisciplinar de conceitos, métodos e teorias presentes nas mais variadas áreas de estudo, salientando-se a teoria de sistemas, a teoria da cibernética aplicada a projeto e os processos algorítmicos, que fornecem as bases necessária para compreender as complexidades da arquitetura algorítmica, com um foco particular em metodologias que possibilitam a integração da modelação paramétrica no âmbito de projeto arquitetónico. Uma área de estudo interdisciplinar que têm as suas teorias e temáticas transpostas para a Arquitetura pelos investigadores e estudiosos do pós Segunda Guerra Mundial, uma época marcada pelo desenvolvimento tecnológico que permitiu o surgimento das metodologias de trabalho computacional atualmente praticadas na arquitetura e do atual computador, ferramentas de trabalho indispensável em qualquer atelier. Além disso, a Dissertação explora algumas das possibilidades do uso das emergentes ferramentas de Inteligência Artificial com base em algoritmos gerativos, particularmente com o objetivo de gerar imagens, de certo modo automatizado, para fins de apresentação e comunicação do projeto ou exploração de ideias não consolidadas. Discute-se deste modo, o potencial da IA no processo criativo, com uma breve analise de alguns dos desafios e considerações éticas associados à adoção destas ferramentas. O estudo apresenta uma visão, que pretende demonstrar tanto as oportunidades como as limitações desta nova ferramenta, bem como a sua utilidade para a arquitetura. Adicionalmente e como referido anteriormente, este trabalho assenta-se numa componente pratica com base em experiências de trabalho pessoais do autor durante a sua participação no atelier Tomás Taveira e Associados. Partindo da análise de cinco casos de estudo, que correspondem a situações onde o problema arquitetónico foi resolvido por processos de trabalho convencionalmente praticados no atelier, mas que revelaram dificuldades e limitações no processo para a obtenção do resultado e/ou no resultado obtido. Deste modo, foram propostos processos alternativos que pretendem colmatar as fraquezas dos processos utilizados pelo atelier. Em conclusão, esta dissertação destaca a importância de uma avaliação equilibrada e crítica das novas tecnologias na arquitetura, em particular a modelação paramétrica e a IA. Um trabalho que não pretende convencer aqueles que o leem a adotar de uma forma acrítica estas ferramentas, mas sim motivar a procura de novas possibilidades de trabalho em que as conclusões sobre as mesmas devem resultar da experiência real e prática avaliada em contexto, e não de especulações baseadas em preconceitos, por vezes arrogantes, sobre o desconhecido.
This Master's Dissertation investigates the potential of parametric modeling and generative artificial intelligence (AI) in the context of architectural project and design, emphasizing the integration of these computational work processes as tools to support architects in a studio environment. The work is developed with a theoretical component of literature research and consolidated knowledge, and a practical component of testing alternative work processes based on the author's work experiences in a studio environment. The theoretical component explores the applicability and value of design work based on computational thinking, which, although one of the foundational pillars of Computer Science, is not directly tied to the use of computers. Considering that computation is the study of “problem solving” processes, involving the multidisciplinary accumulation of concepts, methods, and theories from various fields of study, particular attention is given to systems theory, cybernetics applied to design, and algorithmic processes, which provide the necessary foundations for understanding the complexities of algorithmic architecture. There is a particular focus on methodologies that enable the integration of parametric modeling within architectural design. This is an interdisciplinary field of study whose theories and topics were brought into Architecture by post-World War II researchers and scholars, a period marked by technological development that allowed for the emergence of the computational work methodologies currently practiced in architecture, as well as the modern computer, indispensable tools in any studio. Furthermore, the Dissertation explores some of the possibilities of using emerging Artificial Intelligence tools based on generative algorithms, particularly for the purpose of generating images, in a somewhat automated manner, for project presentation and communication or the exploration of unrefined ideas. The potential of AI in the creative process is discussed, with a brief analysis of some of the challenges and ethical considerations associated with the adoption of these tools. The study presents a perspective aimed at demonstrating both the opportunities and limitations of this new tool, as well as its usefulness for architecture. Additionally, as mentioned earlier, this work is based on a practical component drawn from the author's personal work experiences during participation in the Tomás Taveira and Associates studio. Based on the analysis of five case studies, which correspond to situations where the architectural problem was solved through conventionally practiced studio methods but revealed difficulties and limitations either in the process of obtaining the result or in the result itself. Consequently, alternative processes were proposed to address the weaknesses of the methods used by the studio. In conclusion, this dissertation highlights the importance of a balanced and critical assessment of new technologies in architecture, particularly parametric modeling and AI. It is a work that does not aim to convince its readers to adopt these tools uncritically, but rather to motivate the search for new work possibilities, where conclusions about these tools should result from real and practical experiences evaluated in context, rather than from speculations based on sometimes arrogant prejudices about the unknown.
This Master's Dissertation investigates the potential of parametric modeling and generative artificial intelligence (AI) in the context of architectural project and design, emphasizing the integration of these computational work processes as tools to support architects in a studio environment. The work is developed with a theoretical component of literature research and consolidated knowledge, and a practical component of testing alternative work processes based on the author's work experiences in a studio environment. The theoretical component explores the applicability and value of design work based on computational thinking, which, although one of the foundational pillars of Computer Science, is not directly tied to the use of computers. Considering that computation is the study of “problem solving” processes, involving the multidisciplinary accumulation of concepts, methods, and theories from various fields of study, particular attention is given to systems theory, cybernetics applied to design, and algorithmic processes, which provide the necessary foundations for understanding the complexities of algorithmic architecture. There is a particular focus on methodologies that enable the integration of parametric modeling within architectural design. This is an interdisciplinary field of study whose theories and topics were brought into Architecture by post-World War II researchers and scholars, a period marked by technological development that allowed for the emergence of the computational work methodologies currently practiced in architecture, as well as the modern computer, indispensable tools in any studio. Furthermore, the Dissertation explores some of the possibilities of using emerging Artificial Intelligence tools based on generative algorithms, particularly for the purpose of generating images, in a somewhat automated manner, for project presentation and communication or the exploration of unrefined ideas. The potential of AI in the creative process is discussed, with a brief analysis of some of the challenges and ethical considerations associated with the adoption of these tools. The study presents a perspective aimed at demonstrating both the opportunities and limitations of this new tool, as well as its usefulness for architecture. Additionally, as mentioned earlier, this work is based on a practical component drawn from the author's personal work experiences during participation in the Tomás Taveira and Associates studio. Based on the analysis of five case studies, which correspond to situations where the architectural problem was solved through conventionally practiced studio methods but revealed difficulties and limitations either in the process of obtaining the result or in the result itself. Consequently, alternative processes were proposed to address the weaknesses of the methods used by the studio. In conclusion, this dissertation highlights the importance of a balanced and critical assessment of new technologies in architecture, particularly parametric modeling and AI. It is a work that does not aim to convince its readers to adopt these tools uncritically, but rather to motivate the search for new work possibilities, where conclusions about these tools should result from real and practical experiences evaluated in context, rather than from speculations based on sometimes arrogant prejudices about the unknown.
Descrição
Dissertação de Mestrado, Arquitetura, 2024, Universidade de Lisboa, Faculdade de Arquitetura
Palavras-chave
Computational thinking Algorithmic architecture CAD (Computer Aided Design) Parametric modeling AI-driven image generation Pensamento computacional Arquitetura algorítmica CAD (Computer Aided Design) Modelação paramétrica Geração de imagens por IA