“one learning algorithm” hypothesis
CloseDescendants
    Properties