Publication

Capsule networks, but not convolutional networks explain global configurational visual effects