Keynote Lecture: Artificial Intelligence, Big Data, and Study of Culture
Dr. Lev Manovich, City University of New York
During the last fifteen years, a growing number of researchers in many disciplines ranging from computer science and AI to urban studies, communication, and digital humanities have started to use big data methods for the study of culture. This work has generated many fascinating insights but also led to many questions. In my talk, I want to discuss what I see as some of the biggest challenges in looking at culture with computers, and think about the ways to address them. Why do we often approach cultural data using ideas developed in the 18th and 19th century, before digital computers and big data? Can we think about cultures without using categories? What does AI can see in cultural artifacts, and what it remains blind to? The presentation draws on my recent book “Cultural Analytics” published by The MIT Press in October 2020.
Session 1: An Introduction into our world through the eyes of artificial intelligence
Sophia Brueckner, U-M Penny Stamps School of Art & Design
Catherine Griffiths, U-M Taubman College of Architecture and Urban Planning
Chad Jenkins, Michigan Robotics
Even as these lines are written, AI is already a massive presence in our lives. Be it in the learning algorithms that suggest products on Amazon, intelligent thermostats and voice recognition from Nest, Siri and Alexa. How does the fact that learning and intelligent machines are becoming part of the process of construction, maintenance and everyday life change our habits towards the built environment? When machines have become part of our domestic life, how do they influence social behavior, economic strata and the subjective relationship to the world at large? What is an ethical approach to the use of machine learning algorithms?
Session 2: Do Machines dream of architecture?
Phillip Bernstein, Yale School of Architecture
Matias del Campo, U-M Taubman College of Architecture and Urban Planning
Benjamin Ennemoser, Texas A&M College of Architecture
This session is entirely dedicated to the ways of how ideas on perception changed through the advent of machine vision, and how these new insights into the world have changed methods of architectural design. Machine Vision pertains to an area of machine learning that is preoccupied with methods to inform machines about their environment. Applications of this can be found for example in self-driving cars, recycling centers, and automated visual inspection in fabrication. The ability of Machine Vision to process gigantic databases of images and learn specific features in the process has yielded ideas on the interrogation of existing image material to interrogate them for novel solutions in architecture. This approach is currently being developed in various collaborations One specific area of conversation in this session is the interdisciplinary nature of this new endeavor in architecture.
The ability of learning neural networks allows large amounts of data to be processed in a short time. This property of processing big data as a tool is already being used massively today. The online Google software Deep Dream allows, for example, photos and their styles to be combined. Our increasing understanding of neurological processes has allowed humanity to build machines that literally dream. Do machines dream of architecture? The emergence of Diffusion models such as Midjourny, Stable Diffusion, and Dalle_E2 poses questions to the discipline about authorship and agency within a creative field in which more than 70% of the content is created by algorithms. If machines dream of architecture, what are the memories (the data) that fuels those dreams? In this session, one of the main points of debate is the role of humans in the process of designing architecture with the aid of AI. To this extent, is astonishing that the application of an artificial intelligence algorithm is triggering a renaissance of the question of style. Considering the long overdue un-shelving of the problem of style in architecture, it seems only natural that the current AI research is in the forefront of interrogating aspects of style (Style Transfer, StyleGAN et al). This contemporary conversation can contribute deeply to a novel understanding of style as an expression of our contemporary age. The result of this application is an architecture that is both familiar and strange. A design language that is partially accessible and alienating at the same time. Perhaps the first genuine architecture of the 21st century.
Session 3: Neural Architecture – A paradigm shift in architecture design
Daniel Bolojan, Florida Atlantic University
Matias del Campo, U-M Taubman College of Architecture and Urban Planning
Immanuel Koh, Singapore University of Technology and Design
Sandra Manninger, NYIT University
Kyle Steinfeld, University of California, Berkeley
Neural Architecture embraces the possibility of a design method that is deeply informed by existing information in the form of databases and understands that the artificial modeling of neural processes can aid in harnessing the information in Big Data. Interestingly the results do not resemble historical examples, and thus the methodology is not a repetition of Postmodern tropes, such as collage, quote, and ironic assemblies. The results instead construct a frame around aspects of defamiliarization and estrangement, in that we can recognize certain features without it being a copy. We are still at the very beginning of this new paradigm of architectural production; the first built examples are currently emerging. Many specific problems need to be solved, such as the relationship between interior and exterior, as Neural Networks based on images can only solve one of the two at a time. The problem with bias in datasets, or the issue of utilizing 3D models in this paradigm. Solutions are under development as these lines are written, and I look forward to how this new paradigm will impact the built environment in the upcoming years.
Based on the idea that architecture experienced its first digital turn between 1992 and 2012, the follow-up inquiry is based on the idea that digital tools are no longer in the foreground, but rather the cultural performance that they represent. It almost seems that through miniaturization and cloud services large chunks of technology start to dissipate from the visual focus, the center of the home, and move into the periphery of the household. For architecture, the question arises of its phenomenological qualities. How can AI contribute to the experience of architecture in this environment? Is it part of the materialization of architecture?
Session 4: Roundtable: The emergence of a posthuman design ecology
The Posthuman in this frame of conversation pertains to the idea of shared agency in the design process. It does not mean “after humanity” at all, but rather after the humanist project as for example described by Rudolf Wittkower in Architectural Principles in the Age of Humanism. The discussion in this roundtable pertains to the nature of agency in a design ecology in which human ingenuity meets Artificial Intelligence. What is changed through this novel nature of the design process? Is it, as Mark Burry has pointed out rather an Extended Intelligence, or does AI create genuinely novel architectural solutions? Novel measured by which standards? This roundtable interrogates aspects of sensibility -the ability of artists and designers to invent genuinely new ideas in terms of shape, content, and materialization- and compares them to solutions found by Neural Networks and other AI applications.
The doghouse - ai & architecture 