The Core Principle of Human Intelligence and the Heuristics to Machine Intelligence

The rapid fluctuation of time and space in the quantum world enables the matter forms in the macrocosm to exhibit complex attributes, which in turn enable our brain to have the abilities of representation, memory, and behavior. In three-dimensional space, the neurons interconnect readily with each other via geometric-topological constructions, then this forms the basis of our consciousness and also supports the principle of material primacy in dialectical materialism. We think that the subjective initiative constitutes the essence of human intelligence, which has shaped our learning and thinking paradigms, and also catalyzes human natural language which naturally supports our subjective initiative and guides our reasoning and planning behaviors. The representation model with the perception-array based on the radial basis function principle, or termed the convolutional principle, is widely used to represent color, voice, shape, smell and taste, even the inspired neural network in deep learning, and accordingly we have done an effective estimate of the maximum number of distinguishable color attributes perceivable by humans. It is known that the stimulation response of the simple neuron in the primary cortex is sensitive to the line-orientation, we further analyze the brain's representation mechanism for a geometric shape which is also the representation models of natural language and mathematical language. We have known that there exist two different brain regions, that respectively analyze what an object is and where it is, so any basic semantics-cluster such as a word or a sentence is represented through a unique neuronal fiber plexus node through the layer-by-layer mapping of an underlying geometric topology just analogous to the node in the neural network in deep learning technique. According to the core concept of category theory, all the semantics have been defined by an elaborate semantics bundle structure in our brain, and we can use them to simulate and interpret the physical world including our behavior. Just as the mathematician Poincare's comment on category theory, it is an art of naming, and we accordingly derive the name-abstract mechanism that has enhanced greatly our abilities of reasoning and planning. Thus, when we perceive a semantics scene, our attention will activate a complex semantics-ripple not a simple conditional reflex, and a behavioral closed-loop termed an intelligence-flywheel can realize the iteration of a semantics bundle structure. We explain why our brain consumes much less power. In the end, from the human intelligence principle, the machine intelligence based on more advanced computer database technique with the semantics-driven or termed the structure-driven mechanism can overcome the bottleneck of von Neumann computer architecture in simulating intelligence.