About machine learning

About machine learning

Every learning process is basing the information, what is stored in memory. And when we are trying to teach something to the computer, we are facing the difference between human brains and computer. If we want to drive a driver's license, we have a huge database for supporting that process, and the thing is that we must not learn everything from the empty base. The thing is that at the age of 18, we must not learn things, like what is the barrier, or should we avoid things like traffic lights.

Those things are clear for human beings, and if something is not clear, we must ask that thing, or our instructor just says, what we must do when we are facing some stop sign. And this is the thing, what causes accidents in robot cars. Computers know only things, what are programmed in their memory. But when the robot car is traveling on the road, the problem would be that there is a huge mass of information, what this system must handle.

The thing is that every each traffic mark, or things what the car faces must be programmed in the memory, and then the computer would compile the information, what comes from the cameras and other sensors to the things, what are stored in the memory of the computer. And this is the problem. The thing, what makes computers great calculators is that they handle data precisely, and there are no inaccuracies in that process. That thing makes the data handling process of the car very difficult because if the traffic mark is dirty, the system would not recognize that thing.

And that causes problems. Also, special things like some road works or the cases that the car would face some unexpected things like cable ferry would make very big problems, because if the car faces the sign, what it doesn't recognize, would it stop, and wait for the new instructions. But what if there are no traffic lights in the cable ferry. The robot car cannot solve the problem, because there is no information, what to do in the situation, that some person in an orange suit shows the allowance to drive in cable ferry.

If we would teach the machine, how to act in the situation, where the person would give allowance to drive in the ferry, we might forget, what happens if some person, who has similar work clothes would raise the hand in the front of the street work and move the wrist for giving orders for the digger. In the worst scenario, that thing causes the situation, that our robot car would just push the gas pedal, and then we are in the pit. The thing is that robot cars are working perfectly when they are operating in the closed tracks is of course very big thing, but it would be the long route, that we can make the car, what allows us to read morning magazine behind the control wheel. But there are already been experiments, where robot cars are operating in their own track and limited areas.