Upcoming events

DocLab Live: Bloodless – Guided by the Ghost of a Korean Sex Worker 25 Nov at 17:00

Special presentation on VR installation Bloodless, in which director Gina Kim and guests discuss the influence of American military presence on the South Korean people since 1956.

more events

Learning to See: Hello World!

An interactive installation that shows how machine learning works by rapidly comparing patterns and distilling an image from them.

When activated, an artificial deep neural network is like a newborn baby. Just as the new human brain is fitted with millions of synapses, an artificial intelligence that has just seen the light of day already has a complete internal architecture. But at the moment that it “opens its eyes,” it doesn’t have any frame of reference yet. Every piece of information it receives is new and unintelligible. But the system learns with incredible speed by recognizing patterns in what it observes and comparing all new forms of input with what it has seen before. In this way, it expands its worldview in the hope of contextualizing future observations. With the help of live input from cameras, artist Memo Akten investigates how machine learning works. The way in which the computer trains itself sheds light on the development of the human brain. But when does learning become understanding? Is this a purely human capability, or is it something computers will one day also master?

Project details

Year of development :2017
Created by:Memo Akten

Available formats