Learning intuitive physics and one-shot imitation using state-action-prediction self-organizing maps