There are two ways to make AI. Another way to make artificial intelligence is programming. And another is machine-based AI. Machine-based AI is made by using artificial neurons. The artificial neuron can be the microchip that has its mass memory. That mass memory has its database or table. In that table is stored the thing that the system must know. And the thing that the system must do when it faces the thing. Artificial intelligence is the group of databases.
Machine-based artificial intelligence is similar to program-based artificial intelligence. But in machine-based artificial intelligence, every table has its individual computer. When the computer finds a suitable solution it stores it in the database. And the learning happens by making new databases but also interconnecting databases. The reason why we cannot make artificial brains is that there is needed 200 billion microchips.
One of the most futuristic versions of machine-based artificial intelligence is the atomic or ion cloud. The quantum fields and the quantum strings are acting like the membrane of the neurons. This thing can be the next step in the quantum annealing supercomputers. That kind of quantum computer can learn things spontaneously.
The program-based artificial intelligence means that the programmer makes the network of databases. Which is includes the actions and reactions of the cases. An example of action is that the robot walks on the street. Suddenly robot notices the lamp post and changes its direction for avoiding impact.
So the robot notices the block and makes the evasion maneuver. The problem with program-based artificial intelligence is that if the operational area of the artificial intelligence is very large that makes the program very complicated and heavy to run.
Making the program code for the robot that walks on the streets or drives a car in cities is more complicated than making program code for some military robot submarines. In the case of the military submarine, there is needed only the code that if there is peacetime avoid the contact with other ships.
And of course, parameters that order the submarine to keep a safe distance from underwater mountains and oceans floor are important. In wartime, the submarine needs only the targets that are allowed. Of course, there must be programmed maneuvers for the cases of sudden attacks and something like that. But routine operations for that thing are not very complicated. The number of variables is the thing that determines the complication of the programming code.
But when the robot is driving the car. There are more variables than in the case of the nuclear submarine. The robot must know the rules of the traffic. And if the robot walks to shop and buys things that robot needs more parameters. The system must recognize the things that it takes. Of course, it can make that thing by checking QR-codes and then it can ask the product from the internet by using the computer of the shopping center.
In that case, the robot outsources some actions to outside computers. When the master tells the robot that it must take apples the robot will connect itself to the computer of the shopping center that is telling the location of apples. And the surveillance camera can observe the actions that the robot does.
And give instructions for the robot on what it should take. The robot must also know which force it uses for taking the eggs or open doors. If those parameters are not well-done the robot destroys eggs and everything around it.
When the battery of the robot is low. That system must find the socket and sit for a while to recharge the battery.
The learning system would benefit the records of the past shopping. So the type of apples must not determine all the time. The system might ask about the favorite taste by using the form. Or it can send the test sample of the apples and then the person must just show which was the best.
Then the system knows what kind of taste the customer has. If that precise apple is not for sale the system would select a similar apple. If in its database is the descriptions of the taste of apples. And of course payments and other things like being polite and avoiding impacts must determine for the system.
https://phys.org/news/2021-10-breakthrough-proof-path-quantum-ai.html
Image:https://images.financialexpress.com/2021/03/artificial-intelligence.jpg
Comments
Post a Comment