有教师的学习：前提是要有输入数据及一定条件下的输出数据，网络根据输入输

Teacher's learning: The premise is to have input data and output data under certain conditions. The network adjusts its weight according to the input and output data, so the purpose of learning is to reduce the error between the network's expected output and the actual output. When the error reaches When the range is allowed, the weight will no longer be changed. <br>(2) Learning without a teacher: Only input data is provided without corresponding output data. The network checks the law or trend of the input data, and adjusts the weight according to the function of the network itself. The learning process is: the system provides dynamic input signals, so that each unit competes in a certain way, the winner's neuron or its neighborhood is strengthened, and other neurons are further suppressed, thereby dividing the signal space into multiple useful areas . <br>(3) Reinforcement learning: This kind of learning is between the above two kinds of learning. The external environment only gives evaluation information (rewards or punishments) to the output of the system, and does not provide correct answers. The learning system improves its own performance by strengthening those rewarded actions

There is a teacher's study: the premise is to have input data and under certain conditions of output data, the network according to the input and output data to adjust its own weight, so the purpose of learning is to reduce the network should have the output and the actual output between the error, when the error reaches the allowable range, the weight will no longer be changed.<br>(2) No teacher's learning: only input data is provided, no corresponding output data. The network checks the laws or trends of the input data and adjusts the weights according to the function of the network itself. The learning process is that the system provides dynamic input signals that allow the units to compete in some way, the winner's neurons or their neighbors are strengthened, and other neurons are further suppressed, dividing the signal space into useful regions.<br>(3) Intensive learning: this kind of learning is somewhere between the above two kinds of learning, the external environment only gives evaluation information (awards or punishments) to the output of the system, and does not provide the correct answer. The learning system improves its performance by reinforcing those actions that are awarded.

Teachers' learning: the premise is to have input data and output data under certain conditions. The network adjusts its own weight according to the input and output data. Therefore, the purpose of learning is to reduce the error between the network output and the actual output. When the error reaches the allowable range, the weight will not be changed.<br>(2) No teacher learning: only input data, no corresponding output data. The network checks the rule or trend of the input data, and adjusts the weight according to the function of the network itself. The learning process is: the dynamic input signals are provided by the system, so that each unit competes in a certain way, the winner's neurons or their neighborhood are strengthened, and other neurons are further suppressed, thus the signal space is divided into several useful regions.<br>(3) Reinforcement learning: this kind of learning is between the above two kinds of learning. The external environment only gives evaluation information (reward or punishment) to the output of the system, but does not provide the correct answer. The learning system improves its performance by reinforcing the actions that are awarded<br>