Quantifying Learnability and Describability of Visual Concepts Emerging in Representation Learning

Open in new window