Working Model of neuron cells and Mathematical Modeling

Next, I went on to get started with AI, and I had a rough understanding of what AI is doing and the research objectives. Continue today.

1: Let's review what we 've mentioned earlier. Artificial Intelligence is a manual way to make machines smarter. Intelligence is a kind of behavior. This behavior is related to the human brain, and the brain is mainly responsible for these behavior for neuron cells. It is to study the working principle of neuron cells in the brain to act on machines, let machines have such intelligent behaviors.

2: In the brain, neurons are mainly concentrated in the cerebral cortex. The cerebral cortex is a region with visual and auditory areas. Each region is connected together. Side view6Layer, each layer has inserted columnar tissue, called a neural column.

3: The working principle of the neural column is to transmit the underlying output layer by layer online, and finally process it in the corresponding functional area. Finally, the neural column is processed in a unified manner to respond to external feelings.

4: Neuron cell structure:

We have learned the structure of neuron cells in detail in high school textbooks. I will not explain it in detail here. I will mainly introduce several parts. Nucleus: each cell has its own nucleus. DC: receives input from other neuron cells. Axon: transmits the feelings of current nerve cells. SYN: There are both at the end of the axon and the end of the collateral, and there are neurotransmitters. The transmission of information between neurons depends on these neurotransmitters. Neuron cells are characterized by multiple inputs and single outputs. The input and output are processed by the nucleus.

5: Creation of neuron cell models. The model is mainly based on multiple input and single output, and some processing is performed between the output and the input. It can be summarized as follows:

Multiple inputsX1, x2... XN, As there are adjustments, the correspondingW1, W2... WnIs weight,BInput manually, and then output. Because it is multi-input and single-output, it is obvious that the output function should be written as follows:

This completes the establishment of a mathematical model for the working mechanism of neuron cells.

6: Varies depending on the specific engineering application.FThere are multiple forms, such as linear functions, piecewise functions,SigmodAnd other functions.