Riken researchers controlled a three-qubit system. And that thing can be a revolution for AI research.

The breakthrough in quantum systems. Riken researchers controlled a three-qubit system in a silicon-based quantum computer. And they also made the error correction for that system. So why is this a fundamental thing? First thing is that the silicon-based quantum system is the thing that is needed, for the small-size quantum computer that operates at room temperature possible. 

There is the possibility to create hybrid systems that use magnets and high pressure to make the superconductor. That hybrid system can operate at room temperature. In that case, the high-power magnetic field will stabilize the wire and remove oscillation away from it. 

The AI uses the models of the brain for being effective. 

Do you know why we have three parts in our brains? The cerebrum has two parts or sections. And cerebellum is the part of the brain system. The odd number of data handling units guarantees that the system will not stuck. 

The AI model system will be stuck if there is an even quantity of data handling units. In that case, two parts of the system can make a different decisions. And if there is even quantity of the units that cause a fight. So in the brain cerebellum is required to choose the better solution if the sections of the cerebrum will get a different solution. 

The working artificial intelligence requires the system must have an odd number of data handling units. That thing denies the stuck of the system. 

But the ability to control the three-qubit system makes also possible to create the next-generation artificial intelligence. In that case, the system uses three qubits for emulating the human brain. The system will multiply data into three similar copies. And then that data will send through those qubits.

The reason why that quantum brain has three lines or qubits is simple. The use of three qubits denies the ability where the votes are tied if the system gets different solutions. Because there is an odd number of qubits. That means another possibility always has superior. 

In the case that. There is even a quantity of qubits that thing causes that system will halt. In that case, the system cannot make decisions. The odd number of qubits guarantees that the system will never halt because it cannot make decisions. Artificial intelligence requires a model where another side always gets superior. 

https://scitechdaily.com/quantum-breakthrough-researchers-demonstrate-full-control-of-a-three-qubit-system/

https://everyactioncausesthereaction.wordpress.com/2022/09/10/riken-researchers-controlled-a-three-qubit-system-and-that-thing-can-be-a-revolution-for-ai-research/

Maybe the human brains are the key to making a more energy-friendly and powerful quantum computer.

The researchers of Riken are testing the free-energy-based theories. For creating models for self-learning neural networks. There is a link to that article below this text. The energy-effective quantum computers are the key to self-operating robots. 

But when we want to make an effective self-learning quantum network. We can connect the energy to data that is driven in the system. And another thing that makes those neural networks effective is that every single part of the intelligent neural network would be intelligent. 

They can share the energy overdose. To other parts of the neural network. And that decreases the need to use electricity. That thing can make those quantum computers more energy-friendly. This makes it possible to make smaller quantum computers. 

In neural networks is many small components that are acting as an entirety. And the most well-known neural system is the human brain. One neuron might not very impressive. But 200 billion neurons are turning that neural system into the most powerful quantum computer in the world. 

https://scitechdaily.com/the-free-energy-principle-explains-the-brain-optimizing-neural-networks-for-efficiency/

Image: https://scitechdaily.com/the-free-energy-principle-explains-the-brain-optimizing-neural-networks-for-efficiency/

The principle of AI is that it must do only things made by programmers in purpose. 

The principle of the developers of the AI is that they don't want to make any kind of rebellious robots. Artificial intelligence must make things that programmers want without any nasty surprises. 

The AI must not have the power and abilities that are possible to make. Those systems must have only the wanted abilities. Same way nuclear weapons must not give the most powerful power that they can create. The nuclear weapon must give the wanted and needed power. The thing that makes AI's safe is that they can operate in limited areas. But self-learning platforms are the thing. That can make those systems the same way multipurpose. And able to operate the same way in the large sector as humans. 

This thing can move to the world of AI. The platforms of AI must do only things that are wanted. But if there is a possibility to make robots that can visit shops for people and paint their houses. That thing can seem very nice. But those robots must have limits. There must be a system that allows robots to resist and make a report if somebody orders it to make something illegal. 

The robot is not the same as the AI.

The AI is a computer program or algorithm that can do many things. The robot is that platform that is making physical operations under the control of the AI. So the robot body itself would not be independent. That physical system can cooperate with supercomputers by using the WLAN systems. 

When developers are creating the AI. They are creating systems that can make things what their controller wants. They are making the segment of which kinds of things the AI must do. There is no AI that can do all possible things. 

Modular AI can do many things. In modular AI many separated algorithms can operate independently. There might be different algorithms that are reserved for the military, law enforcement, and civil work. So the same robot body can make all missions that humans can. And the only limit is what skill segment it can use. 

The fact is that modular AI is making robots operate in many areas. One of the examples is the robot servant. There are certain rooms in the house. And there is a series of actions or certain modules for every room. When the robot is moving to the living room it downloads the module. In that module are the actions that are meant for the living room. 

When the robot moves to the kitchen there might be an electronic calendar. When the robot's owner stores some work in the calendar. That mission is transferred to the AI that controls the robot body. If the owner of the robots will give the order to make the certain food. 

At the first, the robot or controlling AI can search for the receipt of the food from the Internet. The robot would see the freezer if there is the needed raw material. If that raw material is not stored. That robot visits the shop. Putting the required algorithms in modules that denies that the AI's code will not grow too large.