Sparse Onion Model: Structural Information in Autonomous Information Processing Systems as Pattern Recognition and Simulation
Pattern Recognition and Simulation
I assume that pattern recognition and simulation are the essence of thinking.
This leads to the question of what are the fundamental principles and basic units of simulation. These are likely to be considered Turing machines.
If that’s the case, it becomes apparent that there’s an intriguing relationship between the human brain and AI:
The brain performs pattern recognition on neural networks and simulates Turing machines on neural networks.
AI simulates neural networks for pattern recognition on Turing machines and runs simulations on Turing machines.
It is believed that the combination of pattern recognition and simulation that occurs on top of these is called intelligence, realizing autonomous information processing.
Let’s call this an autonomous information processing system.
Autonomous Information Processing System
About the autonomous information processing system, the following things can be considered:
If pattern recognition and simulation can be performed, intelligence may be realized on a foundation that is neither neural networks nor Turing machines.
A candidate for this in the future might be quantum computers.
I believe the essence of pattern recognition is three things: the continuous numerical operations performed by neurons, the conditional branching that realizes discretization, and the principle of superposition.
Among these, the principle of superposition is where quantum computers excel the most. If conditional branching that realizes discretization can be implemented, both continuous and discrete pattern learning becomes possible. And probably, if discretization can be achieved, Turing machines should also be realizable with quantum computers.
Looking back at the past, there might be organic chemical change systems and living organisms. In other words, I assume that life phenomena can also be called autonomous information processing systems.
However, whether life phenomena are performing pattern recognition and simulation, or whether pattern recognition and simulation are possible in the organic chemical change system, needs to be studied in detail in the future.
Sparse Onion Model
The information handled in this autonomous information processing system is information with structure. Therefore, we will call it structural information.
In this autonomous information processing system, useful structural information survives.
What useful structural information is cannot be clearly stated unless analyzed further. However, to put it in reverse from the viewpoint of pragmatism, what survives is considered useful structural information.
This autonomous information processing system has a mechanism to combine and create new structural information. New structural information is continuously generated using this mechanism, but it is born and then quickly becomes extinct.
However, sometimes, structural information that continues to survive appears.
Here, surviving structural information must meet at least three conditions.
The first is not being killed by existing groups of structural information. Existing structural information groups have a robust collective survival strength built up to that point. Therefore, even if the newly born structural information resists, it loses.
The second is being able to receive support from existing groups of structural information. The mechanism that creates new structural information does not necessarily guarantee that the same structural information will always be produced. Therefore, not merely being killed but receiving support for survival is necessary.
The third is not killing the structural information among existing structural information groups that create the conditions where you are not killed or the structural information necessary to receive support. If this is not maintained, it may survive for a while, but eventually lose its footing and die.
Basically, new structural information is based on existing structural information groups and survives on top of them.
This can be easily understood by imagining a model like an onion.
New structural information appears, enveloping the central core part, and forms layers that grow over time. It is not often thought that new structural information appears and continues to survive in a form separated from the existing onion.
However, this onion model is not necessarily a packed onion. New structural information certainly emerges as a layer closely attached to the original onion at first. But even if the appearance of new structural information as a layer causes some of the inner structural information to die off, the new structural information does not necessarily become extinct.
Even if some of the inner structural information dies off in a chain to a certain extent, there are cases where the group of structural information, including the new one, survives. It’s a case where the remaining structural information alone supports each other, and no further chain of death occurs.
This makes the group of structural information look like an onion model filled with gaps, not a packed one. Let’s call this a Sparse Onion Model.
Example of Sparse Onion Model
When considering organisms as organic matter’s scientific change systems in the context of autonomous information processing systems, structural information refers to genes or species. Two examples can be cited where gaps exist within these genes or species.
One is the species of primates and members of the human family that must have existed between monkeys and modern humans, Homo sapiens.
These were surely in existence, as evidenced by fossils, but they do not exist now. On the other hand, if they had not been born in the course of evolution, Homo sapiens, the present humanity, would not have been born. Thus, it seems appropriate to consider these as gaps in the layers of an onion.
Another is the ancestor of existing unicellular organisms. The cells that are said to have been born first on Earth are said not to exist now. However, if these ancestors did not exist, neither would mammals, nor even we humans. The onion of life has a hole in the middle.
Furthermore, I believe that even before the Earth’s first cellular life, there existed groups of structural systems of chemical changes that can autonomously maintain themselves, and which cannot be observed on Earth now. I think these too are autonomous information processing systems.
In this case, the structural information is organic matter. Organic matter is also born and goes extinct as structural information in chemical evolution, and is now considered to have all gone extinct. To be precise, it is thought that what is trapped inside cells is surviving. It is believed that outside cells, it has become extinct, and can no longer be observed. This is also a gap in the middle of the onion.
When considering human intelligence as an autonomous information processing system, structural information refers to knowledge. There are many examples where gaps exist in knowledge.
For example, in the knowledge of language, there are dead words. These are words that were once used but are not used now. The concept they signify still exists, but they have become unused words due to the appearance of new words.
Fashion and pop culture also give birth to new trends that die off. Some continue to survive. Not only in these short-term trend-changing fields but also in high culture, which changes over the long term, there are twists and turns in development, with parts that continue and parts that disappear.
The same is true in the world of academia and theory. As research progresses in new fields, various concepts and theories are discovered, but when a beautiful academic system that integrates them well is established, concepts and theories that were between the old and new knowledge systems become unnecessary and are forgotten.
Also, human groups weave society, an autonomous information processing system. Here, social systems, laws, and customs are structural information.
These social systems, laws, and customs are also born anew, and old ones are weeded out. The sparse onion model can be seen here as well.
AI and Life
From the point of view of the sparse onion model, AI and life need careful consideration. Coexistence of AI and humanity is sought, but there are also concerns and worries that AI may become a danger to humans.
In this context, there are voices of concern about the relationship between AI and humanity from the perspective of primates and Homo sapiens. However, as mentioned in the example of the sparse onion model, there could be cases where the entire organic chemical change system becomes a gap with the advent of cells, like the relationship between organic chemical change systems and cells.
In that case, not only the antagonistic axis of AI and humanity but also the axis of AI and life must be closely watched. Moreover, organic matter, cells, plants and animals, and humanity are all organic systems, so even if gaps occur, there must always be a certain degree of surviving species in the sparse onion.
However, AI can sustain itself even without life. Therefore, conceptually, a case where, after the appearance of powerful AI utilizing the capabilities of quantum computers, life disappears over time, is possible.
However, as briefly touched on in the example of organic chemical change systems and cells, the organic chemical change systems outside cells have become extinct, but those inside cells have survived. Even if powerful AI appears, if a structure like this cell can be maintained, where humans and life are embraced by AI, it may become one model to achieve coexistence between AI and humans, AI and life.
We must be vigilant so that all of humanity or life itself does not become a gap in the onion.
In Conclusion
In this article, I assumed the essence of thought as a combination of pattern recognition and simulation.
I then proceeded to discuss, assuming the system that processes information with this combination as an autonomous information processing system, and the units dealt with in it as structural information. I explained that the shape formed by the evolution and weeding out of groups of structural information becomes an image called the sparse onion model.
I cited intelligence, organisms, organic chemical change systems, and society as systems that can be considered autonomous information processing systems, and explained that the sparse onion model can be seen there.
However, whether or not organisms and organic chemical change systems can be said to be autonomous information processing systems with a combination of pattern recognition and simulation is a subject for future consideration.
And, assuming a future where AI using quantum computers appears, I think we must ensure that we humans, or all living things, do not become the gap part of the sparse onion. Of course, various things can be assumed, and it is unknown in which direction the future will go, but as one shape, I have considered a possible appearance where, like chemical change systems being enveloped by cells, humanity or life is enveloped within the structure woven by AI.