eprintid: 10196976 rev_number: 8 eprint_status: archive userid: 699 dir: disk0/10/19/69/76 datestamp: 2024-09-16 09:56:35 lastmod: 2024-09-16 10:13:14 status_changed: 2024-09-16 09:56:35 type: proceedings_section metadata_visibility: show sword_depositor: 699 creators_name: Lim, Soo Ling creators_name: Bentley, Peter J creators_name: Ishikawa, Fuyuki title: SCAPE: Searching Conceptual Architecture Prompts using Evolution ispublished: pub divisions: UCL divisions: B04 divisions: F48 keywords: Machine learning algorithms, Generative AI, Image colour analysis, Computational modelling, Buildings, Evolutionary computation, Computer architecture note: This version is the author-accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. abstract: Conceptual architecture involves a highly creative exploration of novel ideas, often taken from other disciplines as architects consider radical new forms, materials, textures and colors for buildings. While today's generative AI systems can produce remarkable results, they lack the creativity demonstrated for decades by evolutionary algorithms. SCAPE, our proposed tool, combines evolutionary search with generative AI, enabling users to explore creative and good quality designs inspired by their initial input through a simple point and click interface. SCAPE injects randomness into generative AI, and enables memory, making use of the built-in language skills of GPT -4 to vary prompts via text-based mutation and crossover. We demonstrate that compared to DALL. E 3, SCAPE enables a 67% improvement in image novelty, plus improvements in quality and effectiveness of use; we show that in just three iterations SCAPE has a 24% image novelty increase enabling effective exploration, plus optimization of images by users. We use more than 20 independent architects to assess SCAPE, who provide markedly positive feedback. date: 2024-08-08 date_type: published publisher: IEEE official_url: https://doi.org/10.1109/CEC60901.2024.10612168 oa_status: green full_text_type: other language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 2309493 doi: 10.1109/CEC60901.2024.10612168 isbn_13: 979-8-3503-0836-5 lyricists_name: Bentley, Peter lyricists_id: PJBEN84 actors_name: Bentley, Peter actors_id: PJBEN84 actors_role: owner full_text_status: public pres_type: paper series: IEEE Congress on Evolutionary Computation (CEC) publication: 2024 IEEE Congress on Evolutionary Computation, CEC 2024 - Proceedings volume: 2024 place_of_pub: Yokohama, Japan pagerange: 01-08 event_title: 2024 IEEE Congress on Evolutionary Computation (CEC) event_dates: 30 Jun 2024 - 5 Jul 2024 book_title: 2024 IEEE Congress on Evolutionary Computation, CEC 2024 - Proceedings citation: Lim, Soo Ling; Bentley, Peter J; Ishikawa, Fuyuki; (2024) SCAPE: Searching Conceptual Architecture Prompts using Evolution. In: 2024 IEEE Congress on Evolutionary Computation, CEC 2024 - Proceedings. (pp. 01-08). IEEE: Yokohama, Japan. Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10196976/1/2024086077.pdf