We recommend the new term "super artificial intelligence" to distinguish intelligence simulation developed from deep-structure theory from that of functional theory. Artificial intelligence has long been familiar and customary in computers, robotics, and artificial neural networks; all of these technological products are based on the existing functionalism design ideas. Analysis of the brain nerve based on deep-structure theory is completely different from the physiological analysis based on functionalism. The newborn brain simulation technology developed from deep-structure theory is closer to the real brain in terms of function, and it is more suitable to call it super artificial intelligence.
The basic unit of life adaptation is the vitastate, according to deep-structure theory. All life, including microorganisms, plants, and animals, follow the basic laws of vitastate switching. Each vitastate represents a kind of strategy. Any biotic adaptation can be reduced and understood as the vitastates' response to the various ecological pressures. This is the fountainhead of biological wisdom and is also the explanation for intelligence from the perspective of deep-structure theory. The more complex the vitastate structure, the higher the level of life strategy, hence the more powerful the life intelligence is.
The evolutionary lineage is in line with the upward trend of the complexity of life vitastates and vitastates' configuration from lower organisms to higher. This can be used as the reference point from which to compare the different levels of various organisms' intelligence or wisdom. Although this standard is indistinct and not accurate, it touches upon the essence. Vitastates and deep structure are the bases for the formation and development of life intelligence.
Therefore, the ability to identify the level of biological intelligence from the complexity of vitastates' configuration is a criterion of intelligence assessment. It is of more fundamental value than the behavioral response tests of the existing artificial intelligence. If the competition and design priorities of two kinds of vitastates under two kinds of environmental pressures compare with those of various vitastates under various kinds of environmental pressures, the intelligence of the former is more elementary, and the latter is an intelligence that is more complex. If an organism achieves the flexibility of vitastate competition and optimal use by the higher-level neural centers, which compare and forecast the benefit of different vitastates, this vitastate configuration is more complex and the intelligence is more powerful.
Clearly, the first two neural vitastate structures do not require the higher-level modulation and superposed design. From the sense of efficiency, this kind of modulation design has a fast response without intermediate links and can prompt responses to the risks and challenges. The latter neural vitastates' structure needs higher-level modulation and superposed design. Its intelligence can be used for comparing and forecasting the gains and losses from multiple vitastates in competition under the complex and changeable environmental pressures and can also be used for the design of organization flexibility on the basis of combining vitastates. However, the efficiency is lower because this kind of modulation responds slowly with many intermediate links and repeated comparisons and selections before the final decision, which causes a long delay. It can foresee many potential risks and problems, though. Therefore, once the design is complete, it can reply not only to the present pressure but also to future pressures. This modulation has efficiency in a long-term sense.
The new concept of artificial intelligence based on deep-structure theory contains the following points:
1. The new artificial intelligence is defined as a simulation technique systems based on the principle of brain deep structure, which can conduct their own vitastate modulation, vitastate modification, and new vitastate construction according to the internal and external environment, and which have an intellectual ability equal to or greater than that of human beings. According to this definition, all the simulation technique systems based on the principles of original functional structure, such as automatic machines, computers, artificial neural networks, and so forth, can retain the original title of "artificial intelligence" (AI). Deep-structure simulation technique systems are distinguished with the title of "Super Artificial Intelligence" (SAI) or Super Intelligent Computer (SIC), or Brain Computer.
2. The simulation basis of SAI is the vitastates' modulation of the neural network. The first undertaking is to analyze the vitastate structure of the nervous system by the new methods of adaptational biology and through deep-structure experiments. We can look at the neural vitastate as a meaning unit and then reintegrate the immense knowledge gained about molecules, cells, physiology, tissues, organization, animal behavior, and ecology in order to interpret the neural coding system at the level of the molecule and cell. All these provide the conditions for the simulation of SAI.
3. The design method of SAI is to make a brain vitastates atlas based on the analysis of brain deep structure and then to design a new typical brain simulation model and construct the technological system to improve this design by experimentation. The structure of this system will evolve from simple to complex gradually, and the intelligence will increase continually. The new type of deep-structure computer (DSC) and SAI will finally be formed.
There are three routes to promote the experimental study of SAI: The first is to develop a biotic brain computer with the carrier of real neurons. The second is to develop a simulation brain computer with the carrier of artificial neuron
-based nonliving material. And the third is to directly develop the neural simulation software based on the present computer.
Although these three routes are different, they all require deep structure as their design basis. Along with the brain deep-structure theory, the plan is to crack the specific brain vitastate modulation mechanism by brain deep-structure experiments and to draw the general brain vitastate atlas and the brain neural network modulation atlas.
The third route can be easily realized, but the level of intelligence will be restricted by the hardware. The first route is to duplicate the real brain in the laboratory, but it has three defects: a rigorous environment, weak living material, and the ethics problem. The second route is full of developmental potential. It builds on not only nonliving material hardware but also the deep structure of life. Therefore, it can have either high intelligence or the advantages of a machine. This will be the main direction of SAI in the future.