Reconstructing Unreality

The way our brain constructs a conscious sense of vision is highly complex, and not necessarily a direct depiction of the external world. In other words, everything we see in our thoughts is a reconstruction. Our vision system resolves the fundamental problem of reconstructing a three-dimensional world from two-dimensional retinal images by combining information received from the two eyes. As we increasingly rely on machines to read and make sense of our world, the role of neural networks in reconstructing our environments becomes significant.
This experimental workshop is designed to explore the creative potentials of neural networks for depth estimation. Participants will learn how to train their own monocular depth estimation model, generate depth maps and explore the resulting spatial geometries. We will further explore these depth maps through different visual, real-time immersive and sonic representations.
Jessica In is an architect, designer, creative coder + educator. She is a Lecturer at the Bartlett School of Architecture in London, where she is also a PhD candidate. Her research work explores the role of programming and machine learning in architectural representation. Notions of code, data and drawing for the exploration of immaterial and spatial relationships are at the core of her work.