12.08.2019

Symposium 12 - 14 August, 2019: How Humans and Machines Learn to See. Castle Rauischholzhausen

Schedule

Synopsis

The start of this decade saw unprecedented achievements in computer vision by neural networks using supervised training on millions of ground-truth-labelled images. In the past few years these have been joined by an intriguing new wave of achievements based on
more ecologically plausible learning objectives such as reward, prediction and curiosity. This symposium will bring together machine learning researchers, computational neuroscientists, vision scientists and developmental psychologists to discuss the state-of-the-art in how machines learn to see and reason, and how these relate to human learning.