By Tobias Bolli, Junior Project Manager Academic Relations
Architecture has always been a discipline keen on stretching its boundaries, seamlessly molding together art with cutting-edge technologies. With the arrival of the Information Age, attempts have been made to use Big Data and AI to empower architecture. In our webinar on July 14, 2020 our panelists discussed how these digital technologies open up fundamentally new approaches to architecture.
Prof. Dr. Ludger Hovestadt, Professor of architecture and CAAD at the Institute for Technology in Architecture at ETH Zurich, used Google Books Ngram Viewer (a search engine which charts the frequency of words in texts accessible by google) to highlight an interesting trend. The keyword “machine intelligence” appeared especially frequently in google text corpora around 1980. The keyword “smart building” more or less mirrors this curve. Importantly, “Big Data” dwarves the frequency of both of these terms and started to almost break the chart around 2008. Not surprisingly then, Big Data has been impacting architecture in a major way and continues to be a central theme in Prof. Dr. Hovestadt´s research.
A quick search on google reveals: Big Data is a topic people really care about. Turns out it is of major importance for architecture as well.
Lead by the question what digital technologies mean for architecture, he conducted his research in three phases. Starting out, Prof. Dr. Hovestadt and his team tried to get a grasp of what is going on in computing and develop practical applications for architects - numerous successful spin-offs were the result of that effort. A second phase, which lasted from 2010 to 2018, was dedicated to reflection and theory and to writing publications about the gathered knowledge. In a third and still ongoing phase, Prof. Dr. Hovestadt tries to establish a fresh kind of architectural thinking by taking inspiration from various big data adjacent fields - linguistics and machine learning being just two among them.
According to Prof. Dr. Hovestadt, there is one very uncomfortable truth for architects: algorithms are already powerful enough to design buildings themselves and it turns out they are rather good at it! The more data about a customer´s preference is being fed to them, the better and less machine-like their design will be. Moreover, the algorithms (being derived from mathematics developed between 1880 and 1920) are surprisingly short and straightforward and can cover all sorts of cultural preferences and domains.
A visualization of some of the projects Prof. Dr. Ludger Hovestadt and his team have been working on. The pattern on the top left resembles the “Bird´s Nest” National Stadium in Beijing.
The question then arises: what remains of architecture given the ubiquitous availability of these algorithms and the means to feed them? Prof. Dr. Hovestadt´s answer is remarkable. He thinks we have entered a new renaissance in architecture and that it is imperative for architecture students to become digitally literate: to know how to program and to understand the underlying mathematics. He argues that, now more than ever, we have to understand who we are by drawing on our cultural heritage. This can be done best by contrasting our own culture with other cultures and thus recognizing what makes it unique.
Next, Prof. Dr. Biao Li, Director at the Institute of Architectural Algorithms and Applications at Southeast University, told the audience about his journey to Switzerland where he participated in an academic exchange program in 2004. At ETH Zurich he met Prof. Dr. Bruno Keller who at that time was in charge of the program. Prof. Dr. Keller, he remembers, didn’t approve of his focus on architectural design. “If you can only design then you can do nothing”, was his rather blunt remark. Prof. Dr. Li, far from being discouraged by it, took it as motivation to branch out and join the team of Prof. Dr. Hovestadt.
The panelists from top left to bottom right: Prof. Dr. Biao Li, Prof. Dr. Peng Tang, Prof. Dr. Ludger Hovestadt as well as Libing Gu from swissnex China.
There he learned programing algorithms which come up with surprisingly sophisticated architectural designs. For instance, Prof. Dr. Li used generative design to program a residential house in the traditional Chinese Hui Style. Importantly the program doesn't come up with only one design, it can generate all sorts of modifications, different street characteristics, varied courtyard spaces and so on. With all of them being viable designs, the architect can pick one of them without having to do too much design work him or herself.
Prof. Dr. Li said that architecture can be viewed as consisting of two parts: thinking and design (e.g. structure and construction) as well as description and drawing (e.g. layout and physical models). In the past software merely helped with the second part, namely in describing and drawing buildings. However, today’s software is also capable of doing some of the thinking and design. No longer is software merely a tool among others, it does some of the “mental heavy lifting” too and should thus almost be viewed as a coworker.
Whereas programs in the past have been about assisting architects with description and drawing (tasks which don’t require much intelligence or creativity), today’s programs are powerful enough to do what seemed to be a uniquely human activity: designing architecture.
During the Q&A session Prof. Dr. Hovestadt repeated how important it is to take AI and Big Data seriously. Provided we manage to implement it in the right way, he is optimistic that digital architectonics - far from being something soulless and generic - will help to preserve our cultural heritage. According to Prof. Dr. Hovestadt we are still very much in the early phase of this paradigm shift and have just started out on an exciting path the reinvent architecture.
Given the powerful nature of these algorithms, their human-like ability to come up with architecture on their own, it is natural to ask how they will impact the jobs of architects in the future. Prof. Dr. Li acknowledged that digital technologies will alter traditional job roles. He stressed, however, that architects won’t lose their jobs because of that; rather it will free them up to pursue more interesting tasks.
We want to thank both Prof. Dr. Ludger Hovestadt and Prof. Dr. Biao Li for their time and for introducing us to their forward-looking research. Last but not least, we thank Prof. Dr. Peng Tang for the professional moderation of the event.
Please find a link to the slides and webinar recording below:
