Summer Show 2026
unit-code
The Galliera Museum of Porticoes presents an architecture of AI co-authorship through forensic scanning of Bologna’s porticoes, the city’s defining framing device. Designed using a customised digital latent space, the project establishes a framework for curating portico fragments and retelling their histories, when agency is shared between the designer’s intention and AI interpretation.
The museum combines exhibition spaces, artefact and film restoration facilities, embedding a catalogue of porticoes throughout its architecture. Their spatial and semantic logics are translated into stone fabrication and texturing strategies, blending vernacular spolia reuse cultures with contemporary construction methods.
The project reinterprets Aldo Rossi’s concept of the architect as a memory bearer by positioning AI as an active participant in architectural remembering. It proposes a novel methodology where city, architect, and machine engage in continuous dialogue, translating collective memory into layered material expressions.
This latent space framework ultimately transforms how we interpret big data, unlocking the ability to synthesise new urban forms across any metropolitan context.
Bologna’s porticoes are re-curated through a latent space, where evolving metadata and AI interpretation dictate semantic relationships. The architect navigates between tested operations and surprise, creating a dynamic, co-authored curation system.
The curved atrium fits within the interstitial space between the medieval archaeological ruins and existing bus terminal, introducing a new layer of circulation that connects the infrastructural, civic, historical and recreational programmes on site.
The scheme links the vertical layers of the site by integrating subterranean archaeological ruins, canal, street-level infrastructure, roofscape, and observation tower into the circulation. This extends urban dynamics to various levels of the city.
The quoted porticoes are expressed in various materialities, forms, and scales throughout the building; thereby, the exhibits are curated, reconfigured and ultimately embedded within the architecture itself, open to inhabitation.
With this co-authored methodology, data science and machine learning meet architectural systems. This latent space framework transforms how we interpret big data, unlocking the ability to synthesise new urban forms across any metropolitan contexts.