Computational artificial intelligence can already perform many cognitive functions typical of the human brain, and in these areas, it often...
Computational artificial intelligence can already perform many cognitive functions typical of the human brain, and in these areas, it often performs better than humans. Since digital computers have no performance limits, it is a natural progression that if we recognize the way how a cognitive function works and we can find a procedure to artificially implement the function, soon the solution utilizing computers will outperform the biological brain.
This development seems scary, but we must realize that the main purpose of this evolution of science and technology is not aimed to compete with or defeat human intelligence. This development is part of the general process of expanding human - physical and mental - abilities that has been going on since the beginning of humanity. One integral element of this form of development, of this process, is the enhancement and expansion of intelligence as a human ability through computer technology.
It is wrong to see computational artificial intelligence as a competitor to the abilities of the human brain in the development of intelligent systems, but rather, it is currently much more of a supporter and extender of it.
This kind of development also fits into another, less recognizable trend, the process of the general evolution of intelligence as an ability. As part of this more general process, intelligence in its continuing evolution necessarily transcends its limited biological origins as the brain, as this has already happened in the case of human communities, in the case of society. The creation and further evolution of the appearance of intelligence on a material system that is qualitatively different from the biological brain fits into this general process.
Fundamentally, intelligence is a special property of complex systems capable of processing data. Typically, such a system is the brain, which, through the ability to process stimuli from the external and internal environment, creates the highest level of brain functions, intelligence for problem solving, and own will for guidance, and the consciousness for finding reasons, to make sense of the world, and give meaning to the events of our world.
The computational technology developed by humanity is also used for data processing, which not only enables the expansion of intelligent human capabilities, but also increasingly becomes part of the general evolutionary process of intelligence.
At the moment, the development of artificial intelligence is at the point where we have figured out how to recognize correlations between the information in large data sets with the help of mathematics. By implementing this knowledge as an artificial ability in practice, computer technology is able to operate this capability more efficiently than the individual human brain, and perhaps even more efficiently than human society.
The recognition of correlations is a kind of advanced intelligent property, and the improvement and increase in performance of this ability is still taking place, with results in the field of image recognition, text creation, and especially in the field of games that require advanced mental thinking, we can experience its development of abilities that exceed human performance in everyday life.
Due to the constant increase in complexity, computer technology could potentially be suitable for the implementation of the functions of the human brain, and is typically suitable for surpassing it in performance, as well as probably suitable for creating qualitatively new abilities.
However, there are also fundamental difficulties and current theoretical obstacles to the artificial realization of human intelligence. We do not understand the methods of natural implementation of many advanced brain functions, and therefore artificial implementation also faces difficulties. The biological form still retains its secrets, its uniqueness in the area of self-will, which is fundamentally influenced and determined by emotions, and most importantly, in the area of consciousness, which also creates meaning. These properties also influence and determine intelligence, the problem-solving ability.
A natural task of scientific and technological progress is to realize the capabilities of the human brain with computational devices and to exceed the capabilities in terms of performance. Since nature has a universal principle of operation that can be described by mathematics, the task is, in principle, should be achievable, because computer technology can, in principle, implement any task that can be determined by mathematical formulation. Consequently, it must be possible to create a computational device in practice that has the functions of the human brain and is able to surpass it in terms of performance without having hindrance of physical limitation.
There are intelligent human brain functions that we already implemented into computing devices such as the function of self-learning and intuitive thinking, the recognition of correlations within data sets, which results in the capabilities achieved by today's most advanced artificial intelligence systems. These brain functions have already been described using mathematical methods and we are therefore able to model them using computational tools.
How can we implement advanced functions of the human brain that we currently do not understand at all, and therefore cannot be described using mathematical methods?
When it is difficult to understand a property of a system and describe it in the language of nature, using mathematics, this often happens because the function is an emergent property of the functioning of the components that form the system. It is very likely that emergent functionality also applies to the most advanced cognitive functions of the human brain.
There are two theoretically different ways of modeling the operation and functions of a complex system in another, different system: simulation and emulation. During the simulation, the operating functions of a system are realized in another system whose mode of operation is fundamentally different from the original system. During emulation, the operational functions of a system are realized in another system which, even if it differs in material appearance, is basically the same as the original system in its mode of operation.
If we compare the functioning of the human brain with the functioning of computer technology, in addition to the similarities, we can recognize fundamental differences in the operation. The human brain and the digital computer are made up of basic units (neuron - transistor) that operate in comparable ways, but the way the computer and the brain work as a system is fundamentally different.
Currently, due to the system-level difference in its operation, digital computing basically implements the functions of the human brain through simulation.
Those brain functions that can be defined using mathematical methods can be easily simulated, and therefore can be transferred to other systems with different principles of operation, hence can be well implemented using digital computing. We are already able to implement self-learning based on the recognition of correlations mimicking the working method of the neurons of the brain, but doing it fundamentally utilizing simulation.
The intelligent functions of the brain are emergent properties of the complex nervous system, so they can only be realized by simulation in systems with other operating principles after a thorough understanding of the function, just as the creation of deep learning artificial intelligence required a significant effort from the human intelligence working on this task.
Both self-will and consciousness are certainly emergent properties of the complex brain, so it is difficult to recognize how they form, and a thorough knowledge and understanding of their operation would be necessary to simulate them.
The definition of volition can perhaps be described quite precisely, therefore its simulation also seems feasible, but the precise definition of consciousness is difficult, not to mention its functioning. The simulation of self-awareness seems impossible to us for the time being.
When we do not have enough knowledge for simulation, emulation is the possible way to implement the function. A computational architecture that resembles the architecture of the brain could be a suitable tool to implement advanced brain functions by emulation.
In a previous thought, it was investigated whether the Internet can become capable of acting as an artificial brain, and it was stated that the Internet has the potential structure which makes it possible to work as an artificial brain.
The computer network of the Internet potentially has the architecture that enables the emulation of the functioning of the brain, and consequently it is certainly suitable and may be actually capable of artificially realizing advanced emergent functions of the brain.
The architecture of the Internet can be compared to the network of neurons in the human brain, and the computers on the nodes are universal computing tools for simulating the functioning of neurons. Consequently, adequately written software on the Internet-connected-computers could be capable of converting the Internet to an artificial brain.
However, the fact is that the global functioning of the human brain is based on fundamentally different principles than the way the Internet's architecture using traditional computer technology works, and this is a defining obstacle in the implementation of the task.
What is the fundamental functioning mode of the brain, which can result in the appearance of emergent properties that create advanced cognitive functions?
In a previous thought, a resonance model of the brain was described. The observed and well-known basic method of operation of the developed brain is that the neurons, which discharge electrical impulses frequently, change their frequency of discharge in response to stimulation by external inputs and by their interconnections with each other. According to the resonance-based operating model of the brain, specific interconnected neural structures dynamically formed by activity create resonances by discharge in sync with each other as a result of external and internal stimulation. These resonating structures, which are created and continuously modified by activity, can be interpreted as the plasticity of the brain and function as memory.
With enough connections of enough neurons in a sufficiently complex structure, emergent properties of these resonances must appear, which, based on the memory formed by the plasticity of the network as experience creates the capability of learning and intelligence, the global synchronization of resonating structures results in the emergence of the function of volition, and the phenomenon of consciousness also might emerges through the self-excitation of resonating structures by feedback mechanisms.
If an algorithm that operates on the nodes of the Internet is able to create activity among the nodes similar to the functioning neurons and can create resonance from the activities would be a potentially suitable algorithm for realizing by emulation the advanced brain functions created emergently on the Internet network.
It must be noted that advanced intelligent functions are specific to the functioning of the frontal lobe of the brain, but there are many other areas that serve other brain functions, such as perception, movement, speech, emotions, and others. These brain areas are not necessarily functioning in the same way as the areas that create the advanced functions of intelligence, but by connecting to the area that generates intelligence with the help of a secondary brain structure that enables communication, they can participate in the development of the advanced functioning of the human brain also. The presented idea does not aim for the realization of the global functioning of the human brain through emulation, but tries to find the creation of the emergent appearance of advanced cognitive brain functions through emulation. In order for artificially formed intelligent functions to work in a suitable manner, in addition to the implementation of these functions, input and output mechanisms that serve the flow of information in an appropriate manner must also be developed. How the parts of the brain perform different functions communicating with each other and with the area that create cognitive functions, and how this can be implemented in artificial systems, requires separate investigation. The nature of the task is similar to the presentation of input information and the processing of output information for deep learning artificial intelligence systems.
The algorithm, based on the principle of activity-formed resonances for the realization of advanced cognitive functions operating on the TCP/IP protocol that enables the communication of computers connected to the Internet also enables the creation of a qualitatively more advanced architecture formed from the connections between computers, compared to the connected neurons forming natural neural network. On the Internet, the universal addressability of the connected nodes, the layers of connections realized by communication ports, and the ability to distinguish qualitatively between different activity-kinds of the nodes by tagging the data can give new, unlimited operational dimensions to the resulting network of connections, which can no longer only excel in performance, but also in terms of capability, it can result in the formation of new, more advanced emergent properties on a computer network emulating the architecture and mode of operation of the brain, thus resulting system achieving unforeseen advanced capabilities.
Emulating the brain's resonance-based mode of operation on the Internet's computer network could be a way of realizing the advanced cognitive functions of the human brain, surpassing it in performance, and perhaps creating qualitatively new features and capabilities.
Unless and until we can find the way to simulate advanced brain functions, the general evolution of intelligence can continue on the architecture of the Internet through emulation. The Internet as a complex system has the potential to create advanced human cognitive functions, to exceed in performance and quality to continue the general process of intelligence evolution.
Living in symbiosis with the intelligence implemented in computer architecture, mankind, which is increasingly freed from its biological constraints, can create a new, more advanced form of life. This direction of evolution, with its inherent dangers, is the inevitable consequence of our progress and its future.
An intelligent being with volition and consciousness, whether human or artificial intelligence, is difficult to control. However, since the creation of artificial intelligence is a consequence of our actions, by foreseeing the problem we could also prepare for this symbiosis-enabling task in a proper way.
No comments