Training computers to recognize dynamic events

A person watching videos that show things opening—a door, a book, curtains, a blooming flower, a yawning dog—easily understands the same type of action is depicted in each clip.