Rohin Shah comments on Problems integrating decision theory and inverse reinforcement learning