Understanding videos, constructing plots learning a visually grounded storyline model from annotated videos