ParentFull threadp1esk·So basically there's zero evidence in the paper that capsule networks require fewer training examples, correct?View on HN