Merging Intelligences

For some, generative AI is already tapping on the door of creativity as measured by human standards. This is the result of crafting algorithms inspired in a whole range of forms paradigms borrowed from cognitive science. As a result it is quite common that algorithms learn by “playing” with each other, “concealing” information, “guessing” it, “rewarding” each other for success etc. The nature in the form intelligence built in this process influences in the nature of the responses that the computer gives back to us. This session we will discuss the types of the creativity brought by these algorithms and the extent where it can be argued that is matching human creativity.

Cristóbal Valenzuela – Runway ML: Looking at Space

Runway is a next-generation creative software for content creation. Runway uses computer vision and an array of novel computer graphics techniques to reimagine what modern creative software can do. We believe that generative machine learning will spark a new revolution in creativity, media, and expression. We are a mix of artists, engineers, and researchers tackling complex problems to invent a new category of software in the visual arts and creative industries.

Cristóbal Valenzuela is a Chilean-born technologist and software developer. He is a co-founder of RunwayML. Previously, he was a researcher at New York University mainly working on the development of ml5.js. His projects have been exhibited in Latin America, Europe, and the US. Including NeurIPS, Santiago Museum of Contemporary Art, ARS Electronica, GAM, ACADIA, Fundación Telefonica, Lollapalooza, NYC Media Lab, New Latin Wave, DOCLAB, Inter-American Development Bank, Stanford University, and New York University.

Immanuel Koh – Architectural Sampling

The spatial and formal conception of architecture, and thus its modes of design perception and representation, directly contributes to its machine-learnability; and consequently, its capacity in leveraging today’s machine learning apparatus for design innovation. If text can be sampled and synthesised in Natural Language Processing, image in Image Processing and sound in Audio Signal Processing, how can architectural forms and spaces be likewise sampled for generating new designs? What is a sampling unit of architectural form? This lecture will endeavour to construct a theoretical and technical framework with the concept of Architectural Sampling. Foundational to this new design theory, is the overcoming of architecture’s own longstanding set of conceptual and perceptual assumptions, namely figure/ground, parts/whole and shapes/grammars; and replaced with one that is ontologically ‘flattened’, ‘resolutional’ and ‘probabilistic’.

Immanuel Koh holds a joint appointment as an assistant professor in Architecture & Sustainable Design (ASD) and Design & Artificial Intelligence (DAI) at the Singapore University of Technology and Design (SUTD), where he now directs Artificial Architecture — a transdisciplinary research laboratory that focuses on the design and development of AI models for predictive urbanism and generative architecture. He studied at the Architectural Association (AA) in London before obtaining his PhD at the École polytechnique fédérale de Lausanne (EPFL) in Switzerland. His doctoral thesis “Architectural Sampling: A Formal Basis for Machine-Learnable Architecture” was nominated for the Best Thesis Prize. Immanuel has taught at the AA, Royal College of Art, Tsinghua University (Beijing), Strelka (Moscow), Angewandte (Vienna), DIA (Bauhaus Dessau), Harvard GSD, UCL Bartlett and many others. His works have been exhibited internationally including V&A Museum, while his publications include Architectural Design (AD) and Design Computing & Cognition (DCC). He has also practiced as a designer at Zaha Hadid Architects, as a programmer at ARUP with Relational Urbanism (London), and as a creative coder at Convergeo (Lausanne) and anOtherArchitect (Berlin). His forthcoming book “Artificial & Architectural Intelligence in Design” (2020) interrogates the epistemological implications of AI on architecture, and vice versa.

Jeffrey Huang – Artificial Swissness

In Artificial Swissness we discuss the notion of “cultural resilience” in cities, and question the role of creative artificial intelligence and deep learning in architecture. Can machines automatically learn the hidden essence of an architecture or a city? Can they go beyond quantifiable data and optimization, capture and generate something as fragile and elusive as “Swissness” in architecture, and provide resistance to imported styles and cultural colonialism? What are the dangers and dark sides when we let the machines loose? In our projects, machines learn to distill the essence of Swiss mountain architecture from 3000+ images, and suggest critical interpretations using Generative Adversarial Networks (GANs). We show that GANs themselves are not innocent agents and provide an opaque causality between data and generated design. Machines represent a strange ally in the architectural design process, automatically capturing the deep visual structure of a locality, yet also, almost accidentally, inserting ambiguities for human interpretation: a ghost in the shell, pointing towards new aesthetics, not of optimization and clean lines but of strange imperfection, flux, and sense of becoming. In our experiments with GANs, implicitly pursuing Swissness as an aesthetic choice, this alliance between humans and machines suggest a new kind of architectural agency by merging contextual relevance and accidental novelty, and pointing towards a new definition of Swiss architecture, unthinkable by humans acting alone.

Jeffrey Huang is the Director of the Institute of Architecture at EPFL (as of May 1, 2020), that comprises 25 laboratories and groups in Architecture and the Sciences of the City. He is also the Founding Director of the Media x Design Lab, and a Full Professor in Architecture and Computer Science at EPFL. He holds a DiplArch from ETH Zurich, and Masters and Doctoral Degrees from Harvard University, where he was awarded the Gerald McCue medal for academic excellence. Prior to EPFL, he was an Associate Professor of Architecture at Harvard Graduate School of Design. He was also a Visiting Professor at Tsinghua University, a Visiting Fellow at Stanford University’s d.school, a Berkman Fellow at Harvard‘s Berkman Klein Center for Internet & Society, and the first Head of the Architecture and Sustainable Design Pillar at the Singapore University of Technology and Design (SUTD). Professor Huang’s research examines the convergence of physical and digital architecture. His recent work investigates new artificial design paradigms, theories of experience design, and algorithmic urbanism.

Register to attend

Image (see below): © 2019 Immanuel Koh. Subsampling Ricardo Bofill’s La Muralla Roja.