Fresh Eyes brings recent developments in machine learning (ML) to bear on generative architectural design. To improve the utility of artificial intelligence as a creative partner for design, we have brought together experts from architectural design practice, ML engineering, and design methods research, and have developed methods for the incorporation of user-generated image-based ML recognition models into the evaluation step of a traditional generative design workflow. This cluster uniquely links the familiar environment of Grasshopper with cloud-hosted models trained using the Tensorflow framework. Participants will train purpose-built image-based models to evaluate candidate design solutions based on a variety of tacit and heretofore un-encapsulatable design criteria of their choosing, such as architectural style, spatial experience, or typological features. Participants will then deploy these models to the cloud, and integrate them into functional generative design systems via API calls.
This approach opens up a range of possible design scenarios; while the definition of specific studies will be left to the discretion of workshop participants, potential studies include: Transferring Visual Style from Architectural Photography, Autoencoding Architectural Massing, 3d Spatial Composition from 2d Isovists, Quantifying Spatial Flow, and Optimising Architectural Visualisations for Social Media Exposure.
Adam Menges is a computer scientist and designer located in San Francisco specializing in AI. He worked at Apple and SendGrid before starting his own company, Lobe. Currently he's creating a visual programming language for neural networks, and past works include generated art that’s traveled to art galleries across the world and AI bots to find targeted news articles for users. He’s excited to help others learn more about this field at the upcoming SG2018.
Samantha Walker is a Senior Structural Engineering Professional at SOM specialising in seismic design. To enhance reconnaissance efforts in the aftermath of the September 19, 2017 earthquake in Mexico, Samantha spearheaded the development of a photo recognition tool that utilises machine learning to identify building damage. She is interested in how machine learning can influence and enhance different aspects of building design, construction and assessment and is looking forward to exploring these topics further at SG2018. Samantha holds professional engineering licenses in both Canada and the United States and a Master of Engineering degree from McGill University.
As Firmwide Emerging Technology Leader, Kat Park directs design technology strategy at SOM. A computer scientist and architect, Kat Park spearheads research initiatives as well as identify opportunities and technology trends to innovate the design process related to the built environment. Prior to SOM and architecture, she was an interdisciplinary software developer and interaction designer at Art Technology Group and MIT Media Lab. Kat Park holds a BS in Computer Science & Engineering and a Master of Architecture degree, both from MIT.
Through his research, Kyle Steinfeld, an Assistant Professor of Architecture at UC Berkeley, seeks to illuminate the dynamic relationship between the practice of design and computational methods, thereby enabling a more inventive, informed, responsive, and responsible practice of architecture. He is the author of Geometric Computation: Foundations for Design, and has published widely. Most recently, he authored a paper for the 2017 ACADIA conference titled Dreams May Come that sets out a concise theory of machine learning as it applies to creative architectural design, and offers a guide to research at the intersection of ML and design tools.