Augmented (mate)Reality - Exploring hybrid materials from machine learning to physical production

Aug 5, 2026

Cities face daily challenges driven by changing climate patterns, requiring new ways to design and build urban environments. Existing city networks often contain obsolete buildings whose structures and materials must be replaced, adapted, or reinforced. Biomaterials offer promising options but raise questions about lifespan and the ability to scale. This work develops an undergraduate design-studio approach that integrates ML-driven text-to-image datasets with biomaterials and digital/physical prototyping for sustainable, data-driven hybrid material systems.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Background

Augmented (mate)Reality: Exploring hybrid materials from machine learning to physical production Yota Adilenidou¹ , Stefania Boccaletti2 ¹ University of Westminster, London, United Kingdom, ² University of Westminster, London, United Kingdom [y.adilenidou@westminster.ac.uk](y.adilenidou@westminster.ac.uk) ; [s.boccaletti@westminster.ac.uk](s.boccaletti@westminster.ac.uk) Abstract. While cities face daily challenges from emerging climate patterns, there is a need to rethink how we live and build our urban environments with a new agenda. Existing city networks are filled with obsolete buildings whose structures and materials require replacement, adaptation, or reinforcement. Biomaterials, which have emerged over the last decades, offer an exciting opportunity while questioning the lifespan of materials and the need for temporary constructs. Although significant research has been conducted, the main issue of scaling remains. Whether used as a growing material on a traditional substrate, as an autonomous bio-membrane, or as a structure attached to existing infrastructure, it is crucial to explore advanced techniques for amalgamating living and non-living matter and integrating past and post-fabrication processes. This paper presents work produced within the framework of an undergraduate design studio focused on applying hybrid materials for sustainable, data-driven design, inspired by machine learning (ML) text-to-image datasets. Keywords: Machine learning, generative design, biomaterials, digital fabrication, climate change 1 Introduction This paper presents a synergy between Artificial Intelligence text-to-image platforms and matter distribution. Material properties and dynamics are derived from an open-source dataset to reimagine patterns, textures, and structural relationships. The methodology is applied to urban skins and existing building facades within the city network. Cities are reshaping their boundaries through deep skins, which serve as thresholds between interior and exterior spaces, envisioned as a potential new public space network. The work presented is the outcome of two design studio modules from a 2nd-year undergraduate program in Architecture and Environmental Design. Based on a data-driven design approach, the aim is to connect macro and micro scales through the application of environmental and contextual information in designing new hybrid material systems. Hybrids are created between organic and inorganic matter through a variety of combinations of conventional materials and biomaterials. Hybrid strategies are introduced to merge traditional techniques—both physical and digital—with advanced ML text-to-image/imageto-image tools, physical prototypes, and material experiments. One aim is to explore new ways of material formation to enhance the performance of structures while attributing sustainability properties to traditional fabrication elements. It is also imperative to test new AI technologies within the educational agenda, question and renegotiate authorship, and set boundaries between copying and creative use and application. 1.1 Background There has been extensive research on biomaterials, accompanied by significant criticism, mostly related to their scalability. Initial attempts to scale involved performative facades and column structures, which tested the integrity, connection to environment, and structural capacity of the materials. From the Photosynthetica Curtain of Ecologic Studio to the Indus Tile by The Bartlett Bio-ID Lab and Shneel Malik, algae have been a prominent biomaterial used for bioremediation. In the first project, algae were combined with bioplastic sheets, while in the Indus Tile, algae were made into a paste and 3D-printed onto a clay pattern that would provide the path for rainwater circulation and detoxification (Malik et al. , 2019) . Pasquero et al. discuss the Urbansphere and human behavior in relation to other species, as well as an architectural s

FAQs

What problem motivates the document’s approach to hybrid materials?

Urban environments require adaptation and renewal of obsolete building materials, while biomaterials must overcome questions about lifespan and scalability under climate-driven needs.

How does the work use machine learning in the design process?

Material properties and spatial patterns are derived from an open-source dataset to reimagine textures and structural relationships, applied to urban skins and building facades.

What educational and authorship goals are discussed beyond material innovation?

The document aims to test new AI technologies within the curriculum, question and renegotiate authorship, and clarify boundaries between copying and creative use in design and fabrication.

Aug 5, 2026

Upload by River Wang