AI Perception and Creativity_Chris Whiteside_8mb_Page_01.png

AD Thesis Final Submission_20033411_Page_02.png

AD Thesis Final Submission_20033411_Page_03.png

AD Thesis Final Submission_20033411_Page_04.png

AD Thesis Final Submission_20033411_Page_05.png

AD Thesis Final Submission_20033411_Page_06.png

AD Thesis Final Submission_20033411_Page_07.png

AD Thesis Final Submission_20033411_Page_08.png

AD Thesis Final Submission_20033411_Page_09.png

AD Thesis Final Submission_20033411_Page_10.png

AD Thesis Final Submission_20033411_Page_11.png

AD Thesis Final Submission_20033411_Page_12.png

AD Thesis Final Submission_20033411_Page_13.png

AD Thesis Final Submission_20033411_Page_14.png

AD Thesis Final Submission_20033411_Page_15.png

AD Thesis Final Submission_20033411_Page_16.png

AD Thesis Final Submission_20033411_Page_17.png

AD Thesis Final Submission_20033411_Page_18.png

AD Thesis Final Submission_20033411_Page_19.png

AD Thesis Final Submission_20033411_Page_20.png

AD Thesis Final Submission_20033411_Page_21.png

AD Thesis Final Submission_20033411_Page_22.png

AD Thesis Final Submission_20033411_Page_23.png

AD Thesis Final Submission_20033411_Page_24.png

AD Thesis Final Submission_20033411_Page_25.png

AD Thesis Final Submission_20033411_Page_26.png

AD Thesis Final Submission_20033411_Page_27.png

AD Thesis Final Submission_20033411_Page_28.png

AD Thesis Final Submission_20033411_Page_29.png

AD Thesis Final Submission_20033411_Page_30.png

AI, Perception & Creativity

BARTLETT, UCL

2020-21

Download PDF

Project Description

Alongside and in support of my final design project, More-Than-Human, my thesis explored the potential of deep neural networks to challenge human perception and augment the creative process of architectural design. The research included my own design experiments with Pix2Pix Generative Adversarial Networks (GANs), and assessed multiple methods of designing and curating architectural datasets, as well as a number of case studies of pioneering neural architects. It was able to prove the possibilities offered by AI-powered sketch tools and speculated on future photo-realistic rendering tools which could be built with this technology. Through its analysis of creativity and perception, it made clear that the future potential of AI in architectural design lies in the close collaboration and feedback loop between human and machine. It showed that machine perception can be used to challenge our biases and synthesise unexpected, novel possibilities, whilst human judgement can set the agenda and assess the value of results.