Can AI reach human intelligence?

“A leading computer scientist says AI may never reach human-level intelligence because it cannot acquire the tacit knowledge that underlies common sense, intuition, and culture. He argues that this limitation could make advanced AI both fundamentally different from humans and difficult to align with human goals. Credit: Shutterstock (ScitechDaily, Why AI May Never Reach Human Intelligence)

“A new analysis argues that AI may never truly think like humans because the most important parts of human intelligence cannot be programmed into machines.” (ScitechDaily, Why AI May Never Reach Human Intelligence)

The biggest difference between AI and human intelligence is this. AI simply collects and connects information. Human intelligence is deeper. There are more connections with data than AI. Human intelligence. It is connected with culture, intuition, and social behavior. 

Can AI reach human intelligence? That is a good question. The second big, and maybe interesting, question is this: do we want a computer that is as intelligent as we are? But before we start thinking about AI. And its intelligence level. We must understand that there are no humans. Who can do everything. Every single person has limits. And even if we are the greatest mathematicians or engineers. We might not be the best chefs in the world. This means that it might be impossible. To create AI that makes everything. We have different professions and different work. When we want to build a house. We call the construction worker.

And then we need an electrician and a plumber. Those three people could theoretically make a house. But here we see that in those everyday projects. We need three different people. So here is the thing. That we must realize. What one individual person can do is more limited than the skills of the entire human race. We can create AI that does everything.  

That humans can as a species. But that kind of AGI (Artificial General Intelligence) requires that the entire human population will be put in singularity with the AI. The AGI is the system. There is the central AI. That central AI shares missions with AI agents. And each AI agent is a tool that has specific functions. This makes it possible to create modular AI that can connect. New skills. Into it. This thing. Is the main element of the AI. 

When we think of thinking. That means the ability to connect data from different sources. We can say that the AI. It's like a savant. It can have a very good knowledge of things in a thin sector. In a thin sector. AI can beat humans. But other things, outside that real thin focus. They are far below average. This means that in certain areas. The AI. It can be better than humans. But those skills are more limited than humans'. The AI's ability to process. Preprocessed and sorted data is incredible. But otherwise, AI has its limits. 

But then we must ask one thing. Do we require too much from the AI? AI is not like a human. It has imagination. That is based on data. It can get and connect data. The simulation is based on natural laws. And data that is stored in the AI's memory. So, AI cannot imagine things like Superman. Who flies faster than light. Humans don’t fly. Except if they have an airplane. But the AI’s imagination is more limited than human imagination. So the AI cannot think abstractly. It can simulate things. But it cannot create abstraction. All AI’s imagination. It happens by connecting data from databases. 

The worst case is ot system. That makes errors. The worst case is the system. That believes that it's right. Even if the answer is wrong. The ability to operate without errors doesn’t make anything intelligent. Intelligence is how the system detects and fixes its mistakes. Same way. As a professional electrician. Is not a person who doesn’t make mistakes. The professional detects and fixes mistakes faster than a non-professional. 

The big thing is that. AI can connect information from different sources. But AI doesn’t think. This makes AI a problematic thing. If some hacker breaks into the AI’s database. And puts the data that Superman is real into that database. The AI thinks that Superman is real. The AI can double-check all data that is loaded into it during the training period. That happens by using two or three sources. Or, simply by asking the trainer. But when data is checked. And it's in a trusted database. That. Can mean. That the system will not make a new check. And that makes it possible that the data that the system uses is corrupted. Same way. If the trainer makes a wrong decision. That can turn fact into fiction. And fiction into fact. 

So. Is AI smarter than we are? The fact is this. AI can be a better chess player. Than humans. The AI can mimic humans. It can give impressive answers. But the fact is this. All those answers that AI gives. They are formed when the AI connects information. The difference between AI and human brains is that every neuron is like a miniature brain. When AI drives information through the system. There is no need to analyze data that travels in the AI. In natural brains. The brain can analyze data. It can make that analysis. When information. It travels between neurons. That makes error detection easier. 

When data travels through a computer. The system. It makes error detection easier. But only. When the information travels through the entire system. Then the error detector. Makes all calculations backward. So, it drives information in the opposite direction. If the value after that process. Is the same. As  the starting value. The calculation should be right. If. human brains make that thing. Every process in the brain takes a very long time. But. Because brains can make. error detection before the end of the process. And without stopping the main cycle. That can make them more effective. 

https://scitechdaily.com/why-ai-may-never-reach-human-intelligence/

Could fast-spinning aerial vehicles be behind some UAP cases?

“The Phantom Twist drone’s unique rotation renders it almost invisible when in flight (Image Credit: Northwestern University).” (The Brief)

Phantom Twist drone turns invisible. Because it spins 25 times per second. That is too rapid for the human eye. That ultra-fast spin makes its structure “invisible” to the naked eye. This type. Of quite low-cost visual stealth systems. It can make those drones very effective. And that raises a very interesting idea. The image is the AI vision of the next text. The rotor could be larger. But you might understand the idea. 

Could that kind of structure be the thing? That is behind the UFOs or some UAPs. The idea is that there could be a bi-layer rotor. And the middle of that rotor could be the chamber. That is mounted to the frame between those layers. That chamber. It has a bearing connection. That allows it to stay stable. Even if that wing rotates. 

Futuristic vision of the rotating virtual saucer. That spins very fast. (Gemini AI)

When that system travels slowly. Or. It hovers above the ground. Those wings rotate very fast. When that system must travel fast. It stops its spin. And then ignites the jet engines on the other side of the craft. The craft can have similar wings. To aircraft. So that it can turn. As in the case of a cruise missile. The system could look like a saucer when the wing rotates. But the thing is that this image. It is only an AI-made image. It is based on my own imagination.  And created by Gemini.

The mysterious UAP located over Pantex plant. (Interesting Engineering)

The thing is that. The UAPs are new things for the news. There has been a star-shaped UAP above a top-secret nuclear base in Texas. Those UAPs can be some kind of security tests. But they are interesting anyway. And I think that this kind of system could be a test for some exotic-looking aerial vehicles. Maybe some of those UAPs are holograms. The purpose is to make people look at them. That thing could uncover some classified aircraft. The star-shaped structure. It can be. Some kind of stealth helicopter or drone. 

“Jellyfish UAP” might be some kind of rescue drone. The mission of that drone. It could be a tool that can also capture targeted persons. 

There are reports. That  small-sized saucer-shaped object. Traveling along the aircraft, like an F/A-18. Could those saucers be some kind of stealth cruise missiles? The downed F-15 pilot told stories about the mysterious “jellyfish” UAP. That UAP could be a drone whose purpose is to pick up the downed pilot into the chamber. And then fly to the hovering VTOL aircraft. The structure of that system. It can be similar to the Ukrainian drone. That used to shoot down Shahed drones. But that jellyfish drone can have manipulators to pick up the pilot inside it. 

Or maybe they are used for some kind of tests for psychological warfare systems. If the sound waves and the flashing holograms can activate the sleeping centers in human brains. That thing. It can be the new type of intelligent weapon system. But as we know, the UAPs are real. Maybe their purpose is simply to make people look away from something that is not meant for the public eye. And those things require a new type of research. If those systems are holograms and decoys. Researchers might want to collect feedback from those observations. 

https://interestingengineering.com/culture/pentagon-ufo-files-texas-nuclear-plant

https://www.msn.com/en-us/news/technology/what-is-a-jellyfish-drone-swarm-and-why-are-they-considered-so-dangerous/ar-AA27KRUK?ocid=BingNewsSerp

https://thedebrief.org/can-you-see-it-this-stealth-drone-deploys-a-human-perception-hack-to-become-nearly-invisible-to-the-eye/

https://www.twz.com/air/what-was-the-jellyfish-like-drone-swarm-the-downed-f-15e-pilot-reportedly-saw-over-iran

AI might think differently than humans.

“York University researchers have uncovered a surprising mismatch between how artificial neural networks and primate brains process visual information. Credit: Shutterstock.” (ScitechDaily, Scientists Discover AI Models May Not Think Like the Brain After All)

AI might not think like a human at all. When researchers try to make computers think like brains. Those people sometimes forget why human brains are so different. Thinking in human brains is more closely connected to physical structures such as neurons and neurotransmitters. Than we even think about. 

This is the big difference between computers and human brains. The structure of the human brain plays a bigger role in the thinking process. Than we thought. The computer that runs billions of databases. It can have a very impressive capacity to connect information. But the problem is that. One binary processor. It can run one operation at a time. The system requires new processors. 

That processor. It can run. Multiple tasks at the same time. This is not possible for the regular processor. But the processor. That mimics the brain. It can use all its layers as independent processors. 

 The brain has four main areas. 

1) Cerebral cortex 

2) Cerebellum

3) Brainstem 

4 Cerebral hemispheres. The last one has two parts. 

The fact is that if we create a microprocessor that mimics the human brain. We need a five-layer microprocessor. Each of those layers can use a different programming language. That denies the data mix in the system. If every layer of the processor uses different languages, that makes data meant for other layers seem white noise.

Each layer mimics each part of the brain. And those layers. They can have different programming languages. Or different frequencies. That helps them to separate and sort information. The four-layer microchip. It can have one divided layer. That mimics the cerebral hemispheres. 

From different sources. The database connections. They can mimic neural networks. That transfer information in human brains. The difference between database structures and human brains is this. Databases run on the same monolithic computer. In human brains. Every single neuron is like an independent computer. This means that if we want to make a computer. 

Or. We can rather say: an AI solution. We must create a system. That involves 86 billon computers. The system. It can use morphing neural networks to make its operations more effective. 

“Schematic of a simple feedforward artificial neural network.” (Wikipedia, Neural network)

The thing that can make the process quite easy is that. Every neuron has. The ability to change its connections and their relations. This means we cannot measure the neuron’s ability to process information in the same way. As we measure a computer’s ability to process information. Human brains store information in the form. 

That is similar to a mosaic. The information. It is stored. In a form that is like pixels. Brains can connect and reshape those pixels freely. And that ultimate flexibility makes human brains so different from machines. In brains, every neuron has a pair. A mirror neuron. The neuron and its mirror. They act like loops or algorithms. Another purpose for mirror neurons. That is. They tell the primary neuron that the information traveled through. 

In human brains.  Multiple points of start. Data processing. At the same time. Brains can spread the operation. They could reserve more neurons for that action. This means that brains concentrate. In another way. Than computers. In brains, in brains. Every neuron acts as an independent computer. And the large number of neurons gives fine-tuning for processes in the brain. 

Computers can also connect data. But that system is far different from humans. In computers. The binary system. It can handle only one task per operation. The AI. That mimics human brains. Must have 86 million physical processors to mimic human processes. The ability to search data and then combine that data with memory. Is the thing that we call thinking. 

We could make a machine. That mimics human reactions. That machine requires a physical platform. That involves the same structure. As human brains. The fact is that. If we want to make a machine. That thinks like humans. We must remember that the system. It is a combination of hardware and software. 

https://scitechdaily.com/scientists-discover-ai-models-may-not-think-like-the-brain-after-all/

https://en.wikipedia.org/wiki/Neural_network

What do a locomotive and a data center have in common? They both face resistance.

Political resistance against data centers. Forces them to find new locations for those systems. 

Data centers face resistance. That is one of the things. That we can see. The main problem with data centers is this. Those systems don’t require permissions. Except for the use of land area. The data company. It can buy a large group of houses. And then turn the database into cloud-based systems. The cloud-based architecture means this. There can be extremely large data centers in the neighborhood, and nobody even knows that they are there. 

When we resist data centers. And the use of AI. We can see similar cases in history. In history, people resisted things like cars. The origin of the car is in the train. The steam engine made it possible to make trains. 

The speed of the first train. It was 21 km/h. The crew of the train was two. Two men could handle very much cargo. The first trains were used in mining areas. They could maintain their speed. All the time. And that removed the crew from logistics. And then people started to think about the possibility. To create the train on wheels. The combustion engine made the car real. And that caused problems with workers. Unlike horses. Cars and trains required mechanics. Those people who worked with those mechanical systems required training. Unlike people who feed horses. 

They said cars would cause unemployment. Cars caused pollution. But the first arguments against cars were that cars take work from cattle workers. The problem with that criticism was simple. It was that the critics were cattle owners. Horses were the most important working “tools” before the car. The big problem with cars was that. They took the place that belonged to horses. first car was slower than a horse. The maximum speed of the car. It was about 15-30km/h. But the car was a machine. It could maintain that speed. All the time. So, the car was more effective. The car didn’t need water or food. And that made tractors and cars suitable. To operate in places. Like Antarctica. If people operated there with horses. That required a lot of food. 

But as we know, people resist data centers for many reasons. The big problem is that. The only thing that measures effectiveness. It is the income money. If people have free time. In their workplace. That is ineffective time. That causes a need to decrease the number of workers in the workplace. That causes unemployment. 

Another thing is that. People resist everything that is new. The ICT area is been like in the position of the stepson in the media. When we think about traditional business. That thing required a lot of workforce. A steel factory sends pollution. So that requires lots of permissions. Of course. Data centers require permission to use land area. The data company can just buy a lot of houses. And then make data centers in them. This kind of solution doesn’t need new buildings. This means that. The data centers don’t need as much political support as traditional factories. 

And the second thing is this. Data centers are primary targets for the enemy in the case of war. This means that things like underground facilities can help data center survivability. 

But the answer. It could be an underground data center. The tunnels are full of supercomputers and are not visible from the ground. The technology. That those tunnels require. They can be the same. That is used for making subway trains. 

The ICT company doesn’t require a workforce in the traditional way. They don’t need raw materials. They need people who make code. There are no psychological or health limits in this work. The ICT company just needs working spaces. And remote work makes it possible to operate data centers from another side of the world. The head coder can do the job. That person can operate and train AI agents to make code. The underground facilities are the answer to the natural problems. Data centers that are deep underground. They can use geothermal heat or miniature nuclear reactors to provide electricity. 

The deep caves. Like exhausted gold mines. They can provide stable and radiation-protected locations. Some of those points are used for neutrino telescopes. But those locations can be suitable places for quantum computers. The quantum computer. It can be in a thermos box.  The isolation layer: A faraday cage and radiation protection. They are between the walls. Of the box. The purpose of those layers is to isolate the qubits. From the outside environment. 

The other place where those data centers. They can be made. Is the ocean floor. Large and complex structures. They can be modules. Dropped to the deep sea. Deep-sea data centers can be operated using robots. Underground and deep-sea positions. They can protect data centers against terror and bomb strikes. 

Rising resistance against data centers. Forces data companies to find new positions for data centers. The deep sea and underground positions. They are effective. But things like orbital data satellites are new tools. They can operate using cloud-based architecture. Data satellite. It’s a similar server. To other servers. The orbital server’s program maintenance. It is similar to other servers. So, the person who does the maintenance work. That person doesn’t need to know. The position of the server. The orbital data center. It can be the belt or chain of data satellites. If one of those satellites is visible from a ground station all the time. That means the maintenance will not see any difference between ground-based data centers. And data satellites.  

The satellite. It can have a heat shield. And the ability to land safely. This means that those satellite swarms can recover their critical components. And that helps to find. If somebody tries to “steal” them. The orbital data center. It is a group of satellites. Those satellites can communicate with each other using lasers and radio communication. The system is similar to Starlink. The backups. And other things can be made into other satellites.

Or ground-based hard disks. When one satellite is jammed. That satellite will be replaced with another satellite. The operations with orbital computer platforms are the same as they are on ground-based systems. The people who operate the computers. They drive system updates into the orbital data servers. As they do for the normal ground-based data centers. The people who update and maintain computers. They must not have access to satellite trajectory controls. 

People who adjust satellite trajectories must not sit in the same room. There the compute operators sit. The same way as in every other data center. The maintenance crew can operate those systems remotely. Just like in every case on Earth. The remote operators don’t need to know where their server is located. They need to know how to make those updates. 

Brains might use different languages in external and internal communication.

Brains might not use natural language for logical reasoning. Or they don’t use language. That we can use in a verbal form. This means that the brain’s internal communication requires a different communication protocol. Than humans use in communication with other people. We can call those protocols languages. But logic doesn’t need a similar language. As we use in speech. Logic needs communication. And a way of communication. It can be different. But we can call that communication “language”. 

We understand natural languages as language. That we speak and write. But nature is full of languages that we cannot use in verbal communication. Things like hormonal and pheromone-based chemical communication. It's very common. Most of the neuron’s mutual communication is a combination of electric and chemical signals. So that means those communications. They are also natural. But the key component in information is this. Brains must have an ability to separate external and internal signals. 

Brains are very complicated structures. They contain many hidden neural networks and hidden layers. We know that only a small part of the brain's internal communication reaches consciousness. 

This raises an idea. That maybe humans use different languages to communicate with other people. Than brains use their internal communication. So could the cerebral cortex. The diencephalon and brainstem use different communication languages for internal communication. 

If the brainstem uses the same language for internal communication as the cerebral cortex. That could disturb the brainstem's internal communication. And that can cause death. The brainstem controls vital functions like blood pressure, pulse, and other things. In some models. The diencephalon has a mission. It is to translate brain signals from the cerebral cortex. To the brainstem. If those three brain layers use the same language. That can be hard. To separate the signals. And that can cause problems. 

When we expand this model. We can say that maybe all brain areas, including the cerebellum, have different languages. The cerebrum and cerebellum have different languages. But brain lobes like the frontal lobes and temporal lobes. They use different types of dialects. That helps the brain separate signals. That come from a certain area. And that helps to keep those signals in order. This is important. When. Brains make decisions. About the reactions to a certain stimulus. Every stimulus requires a different type of response. 

So. The use. Of different languages. In internal and external communication protects brains. 

It’s possible that all human brain areas use different types of language. If we want to use computing as a model. How the brain analyzes information, we must take note of. Computers have three types of programming code. The lower-level programming code. That code connects the hardware to software. The operating system level. And the highest part is the program. All three main layers are made using different types. Of programming language. 

Those three layers are meant for internal computer communication. The network communication uses other protocols and languages. Than the computer internal data flow. But then we can transform that model to the human brain. If human brains use different languages than humans use in communication with other people. That protects the brain from outside effects. In this model, information comes to the brain. through the brain area. That is reserved for a certain sense. When other humans speak. Data travels in the brain lobe. 

That handles those types of signals. Information travels to the Broca lobe. There, those neurons translate that signal into the form. That. The brain understands it. If humans. Use the same language for internal and external communication. That thing causes a situation. That other person can take another person under control. Because external communication is different. Than the brain's internal communication. Brains realize. That data comes from external sources. 

https://mcgovern.mit.edu/2026/07/06/separating-logic-and-language/

https://www.pnas.org/doi/10.1073/pnas.2520095123

https://en.wikipedia.org/wiki/Brainstem

https://en.wikipedia.org/wiki/Human_brain

https://en.wikipedia.org/wiki/Human_brain#See_also

Can there be quantum mechanics without imaginary numbers?

Above: Complex numbers. Imaginary parts are marked by i. 

In the beginning. I must say that it's always possible to calculate a point in 3D spacetime. Very accurately. The movement of the object by introducing . Only the series of coordinates of the path. The object followed. Qw cannot draw a 3D image on a 2D layer. But we can list the points. In a 3D coordinate system, there are three coordinate axes (X,Y,Z).  

That the object’s trajectory followed. This list looks like this: (3,1,1) (3,2,2) (2,3,3), but. We cannot draw that graph. By using one coordinate system. If we could share those coordinates in three parts: (X , Y), (X, Z), and (Y, Z). We could introduce a 3D graph on a 2D layer. By sharing the 3D coordinates in three different coordinate systems. 

“Quantum mechanics has long relied on complex numbers to describe the strange behavior of particles, from tunneling to entanglement. A new analysis suggests that this mathematical language may not be fundamental after all. Credit: Stock” (ScitechDaily, Quantum Mechanics May Not Need Imaginary Numbers After All)

Imaginary numbers play a big role in quantum mechanics. Imaginary numbers are part of complex numbers. The first part of the complex number. It is used to determine the particle’s position in space. An imaginary part determines the particle’s energy level. This is the classical way to determine a particle’s position in 2D space. Now researchers are investigating the possibility of making those calculations using only normal numbers. This kind of possibility is the thing. That could make it easier to make those calculations. There is a possibility. That is, the imaginary part of the complex number is replaced. By using two different calculations. That means the primary and imaginary parts of the number. 

They are simply calculated. In different calculations, but using the same formula. In normal imaginary calculations with complex numbers. The complex number’s real and imaginary parts. Can be calculated separately. All of those calculations. They can be made by using real numbers. The system must just calculate vectors. Those are from different angles. We can imagine a situation. The 3D model of the space is transformed into a series of 2D models. The horizontal and vertical positions can be introduced with two different 2D layers. 

“Explanatory diagram for the research question—is quantum mechanics possible with only real numbers?—and results of the study. Credit: Pedro Barrios Hita, HHU” (ScitechDaily, Quantum Mechanics May Not Need Imaginary Numbers After All)

This means that the researcher can use three regular 2D coordinates. To introduce the object’s position in X,Y,Z axes. The normal X,Y coordinates can be used to improvise the position in the 3D space. The first coordinates are regular X,Y. And the others. They are the X,Z, and Y,Z coordinates. If the system must not draw images. The positions of the objects. They can be introduced very accurately using regular coordinates. X,Y, Z. This is the thing for. The system introduces positions in 3- or more-dimensional spacetime as a list of coordinates. So, we could use this model to introduce the positions. And energy levels in quantum mechanics. The system can use different coordinates for the energy levels and positions. This means that the number of those coordinates. They can be increased without limits. 

https://www.mathsisfun.com/numbers/imaginary-numbers.html

https://en.wikipedia.org/wiki/Imaginary_number

Can the AI learn from its mistakes?

That is a good question. The answer is simple. That depends on how the error or mistake is determined. Determining what the mistake or error is is one of the bottlenecks to creating artificial general intelligence (AGI): how to determine right or wrong? How to describe favorable cases. That the AI should use. And. How to determine the non-favorable cases? The latter are cases that the AI should avoid. In the teaching process. The operator determines and describes the case. And then gives it. Positive (favorable) or negative (non-favorable) values. 

In this text. The main topic is reinforcement learning. There, the system makes something, and the environment. It gives feedback. The feedback. Or. The actor who gives feedback. Determines. If the AI acts right or wrong. And how to determine the values “yes” (Positive)(+)  and “no” (Negative)(-)? 

The system could partially follow Boolean algebra. The chain of positive (+) solutions. It can be conjucted by using “AND”. The “OR” changes the model. And if the model gives a negative (-) value. The system turns to using “NOT,” and then the system. It must change the model. The disjunction happens when there are too many negative values in the series of cases. The negation operation makes the system retake the algorithm. And then try another way, or algorithm, to solve the problem. The problem. It must always be solved by following the rules. 

When we think about the trial-and-error model. That model is effective. But not in all cases. This model is also known as the reinforcement model. Trial-and-error model. It is a good tool for virtual cases. But in cases where the AI must drive a car. That kind of learning solution. That can turn very expensive. There are not many ways. How to react to things the right way. Wrong reaction. It can turn fatal. If the AI driver reacts the wrong way. That can be a very big risk.  

When the car stops at a red light. That is the rule. This instruction is for public safety. But what if somebody tries to rob the car? What if a street gang member  shows a red light or “stop sign” to the car? Trying to rob it? That case is not very common. But those special cases show. How difficult. It is to program the AI. The AI is like a student. That system requires intensive training. The AI trainer must give instructions on what to do. And what not to do. 

In simple cases, the AI uses a limited data type. The AI is very easy to teach. The system requires a description of the favorable case. That case is determined as plus. But then the AI requires determination. About the non-favorable cases. The thing that the AI should not do. That is as important as what the AI should do. The AI should also have value. 

What to do if it doesn’t recognize the case? In a virtual world. The AI. It can make as many mistakes as the user allows. But in real life. When AI controls physical things. There is no room for errors. If the AI controls robot forklifts. Those systems can break lots of merchandise. If they work wrong. If the AI controls vehicles. like cars. And it reacts the wrong way. Results can be devastating. In real traffic, the vehicle has no time to wait and analyze opportunities. 

If we want to use virtual environments. The AI can wait. More information for the entire day. The virtual system. It can have endless time to try again. Or wait for more information. 

The world in the virtual environment. There, the system handles things like numbers. There are only two possible cases. Right (+) or wrong (-). But in cases like traffic, there are also plus-minus (±) cases. When AI controls a car. It can face a situation. That there is an emergency vehicle behind it. The AI can be ordered to drive to the sidewalk. The AI must also have orders that it must not impact people. And those cases. That don’t happen very often. They are the most challenging things for the AI. The AI must be prepared. That somebody tries to rob the car. Or there is an emergency vehicle behind it. In a tight avenue. 

Boolean algebra. 

The AI can learn in three main ways. 

1) Reinforcement learning

“In machine learning and optimal control, reinforcement learning (RL) is concerned with how an intelligent agent should take actions in a dynamic environment in order to maximize a reward signal. Reinforcement learning is one of the three basic machine learning paradigms, alongside supervised learning and unsupervised learning.” (Wikipedia, Reinforcement learning)

2) Supervised learning

“In machine learning, supervised learning (SL) is a paradigm in which an algorithm learns to map input data to a specific output based on example input-output pairs. This process involves training a statistical model on labeled data. Each input is paired with the correct output. The term "supervised" refers to the role of a teacher, or supervisor. Who provides. This training data guides the algorithm. Towards correct predictions. For instance, if you want a model to identify cats in images, supervised learning would involve feeding it many images of cats (inputs) that are explicitly labeled "cat" (outputs).” (Wikipedia, Supervised learning)

3) Unsupervised learning

“Unsupervised learning is a framework in machine learning where, in contrast to supervised learning, algorithms learn patterns exclusively from unlabeled data. Other frameworks. In the spectrum of supervision. Including weak- or semi-supervision, where a small portion of the data is tagged, and self-supervision. Some researchers consider self-supervised learning a form of unsupervised learning” (Wikipedia, Unsupervised learning)

“In supervised learning, the training data is labeled with the expected answers, while in unsupervised learning, the model identifies patterns or structures in unlabeled data.” (Wikipedia, Supervised learning)

“The typical framing of a reinforcement learning (RL) scenario: an agent takes actions in an environment, which is interpreted into a reward and a state representation, which are fed back to the agent.” (Wikipedia, Reinforcement learning) The thing that gives feedback. Like determining whether the case is favorable. Or non-favorable. It can be the human. 

The main problem with the AI and learning system is. How to determine whether the solution is good or bad. The simplest way is to use a human as a controller. When the algorithm ends its operation. Human operators. They select whether the solution is right or wrong. Determination of the desired solutions. It can also be programmed into the algorithm. In the case of stock marketing, desired. Or. A favorable solution could be maximized income. In the series of actions, the algorithm repeats the action. Time after time. The solution that it pursues. That is, maximizing income. 

Stock market analysis is a simple solution for modeling. The rising line, or rising income. It is the positive solution. The decreasing line is the negative thing. 

This type of machine learning is not hard to make. The user must only determine the highest number. That is, in a certain column. That is what the AI should pursue. In a series of cases, the user marks the wanted solutions, or actions. As positive (+) and negative (-). The thing. The algorithm must pursue. It is the highest possible number of positive solutions. 

The user determines the plus and the minus. And the AI tries to take as many points in the plus column. As possible. The AI, or its teacher, just selects the answer. That is marked as plus. The process requires more than one point. And then the AI follows the line. When the line is rising. The AI makes the right (+) solution. When the line decreases, the solution is wrong (-). 

https://vertexestechnology.com/levels-of-ai/

https://en.wikipedia.org/wiki/Boolean_algebra

https://en.wikipedia.org/wiki/Supervised_learning

https://en.wikipedia.org/wiki/Reinforcement_learning

https://en.wikipedia.org/wiki/Unsupervised_learning