Parallels

Marc Da Costa and Matthew Niederhauser

Artwork Description:

“An imagined landscape, then, is a landscape not of being but of becoming: a composition not of objects and surfaces but of movements and stillness, not there to be surveyed but cast in the current of time.” ― Tim Ingold, Imagining Landscapes: Past, Present and Future Parallels is a site-specific, responsive machine learning installation that transforms a square LED wall into a portal for visitors to encounter the world and themselves through the lens of a neural network. Carefully framed in the environment to provide the viewer a pass-through cutout of the surrounding world, a camera on the rear of the screen captures an image each second and transforms it through machine learning models that have been trained on the art historical and visual histories of the installation site, as well as other ambient, environmental sensor data. In so doing, the work enables a visceral and unmediated encounter with the gaze of machine vision as it discerns the surrounding environment, dreaming and reframing it in conversation with those who encounter it. The project seeks to recontextualize emerging machine vision technologies within the landscape and built environment, questioning how new forms of artificial knowledge can be placed within a longue durée of human sensemaking practices. The installation thus draws the sphere of machine vision into conversation with the surrounding architecture, natural features and the historical embeddedness of the viewer, providing an occasion to meditate on a moment when ways of seeing are being reconfigured.