Proteus 3.0: Reinforcement learning for intuitive interaction with digitized matter

Maria Smigielska, Pierre Cutellic

Proteus 3 is an installation, which tries to remain in conversation with its audience and learns from the visitors' engagement to interact with lively, digitized ferromaterial in an exhibition space. As an architectural, room-scale oval oculus, it employs the reinforcement learning techniques to understand better human behaviour during an interaction and making it more meaningful.

__________

Maria Smigielska is an architect and researcher based in Zurich, currently working at DBT, ETH. She is interested in enhancement of potentials for creation of architectural elements, design objects and mixed media installations, by using digital and interactive technologies for encoding and modulating materials properties, custom design and fabrication methods.

CompMonks (Pierre Cutellic), PhD, is an architect and a researcher in technology for architecture based in Switzerland. He currently leads research on interactive and generative design with brain-computer interfaces at ETH Zürich. His artistic work follows a series of design objects and mixed-media installations focused on the power of combining humans with computers.