White eyes and understanding of artificial intelligence hypothesis (Part two)

Speaking on the blog post baby on Mom and Dad know of such a process is out of frequency of occurrence

This is a process, we see the baby uses hearing, sight, sound.

In the eyes of each of us, the world we live in is three-dimensional, but the camera is seeing things that are two-dimensional, because people have two eyes, and therefore probably be able to identify whether an object is a whole, given two computer a distance from a camera, the difference is determined between the object shape can be substantially determined whether an object as a whole.

God gives one of the genes is: identify the integrity of the object

I think in class, what is it merely a picture, let us think of an object as a whole. At first I thought is based on his profile, but things are not as imagined as an object, and a positive look sideways to see the profile is different. Class time, I walk in the street to see the profile of those trees is irregular enough, but I know it is a tree and not something else, after a period of reflection I think it is

Two gene God gives is: general description of the object memory

The reason why is the general, because of irregular objects such as a tree, if only rely on the outline, then, is to identify the object does not come out of this tree. Because each tree, the shape is different, so the whole, the tree has his own characteristics, I can know that it is a tree trunk for the first striped brown slightly for the straight cylindrical object, this shape it is the trunk, then there is usually top of green foliage. Even if the tree was cut down, out of the whole, we are still able to recognize it is a trunk. Trunk skin was also stripped down, we still know it's bark. That memory of these things and holistic recognition can be distinguished, of course, recognize the bark it is built on top of that we have some understanding of the integrity of the object can do. If we do not have this concept was never bark, but aware of the existence of the tree, we see the bark when you can probably infer that this is the tree of the epidermis, but we can not say he is the bark, because we do not have a tree skin concept, just from the thought to know this is the skin of the tree. Conversely, if the beginning is known not know the bark of a tree, then after seeing the tree when it is easy to know previously seen is bark from the integrity of the object (tree) Grilled down.

So now we have to simulate such a process, giving computer camera, microphone and speakers

To simplify this process, the process and want the computer to have a baby like corresponding memory of their mom and dad called it, if we identify the computer in a white background, there is no such background knowledge of training a green ball and took out when the green ball, allowing the computer can tell the process of green ball

Because it is a single camera can not identify whether it is whole, only through the algorithm to determine whether he is whole. So if this time the identification of an object is the overall algorithm it has been achieved, or installed two cameras achieved, whenever there has been a white background green ball, we would say that green ball to him, and I told him was after repeated training, computer green ball thing with the integrity of the sound associated with the little green ball up.

To be continued

Guess you like

Origin www.cnblogs.com/halone/p/11758214.html