So let me just sum up the main points of the topic that we've been discussing. The challenge as I've just said for the visual system is getting around the inverse problem. In any theory of vision, whatever you like, whatever appeals to you and as I say these theories that we've been discussing today are not mutually exclusive. They all have their place In thinking about vision, but you've got to take the inverse problem into account. You've got to explain how whatever the theory is that you like and advocate is resolving, there's no solution to an inverse problem, but it is resolving the quandary that this presents. The major options that we've discussed are vision by feature detection, vision by inference, vision as efficient coding, or vision based specifically on a way of dealing with the inverse problem. All of them have their place, I think the inverse problem and it's resolution is the primary obstacle that is facing vision scientists today. So in this last option, which is a biological option, the goal is not seeing the world the way it really is. That's our intuition, that we look out there, we see the world the way it really is, the goal is biological success. And in biology that means, reproductive success, Determined by Natural Selection. Natural Selection allows things that work to go forward, things that don't work to be eliminated. And I would argue that the visual system as it exists today is really the product of that process. Obviously it's evolved but that the circuitry, the anatomy, the physiology that exists today is really in the business of resolving the inverse problem, not in the business of trying to see the world that really exists out there which machine vision can do. But we as biological entities cannot.