1. The origin of neural networks
We have heard of neural networks for some time now, but in fact they have been around for a while. The birth of neural networks originated from the study of the working mechanism of the brain. Early biologists used neural networks to model the brain. Machine learning scholars used neural networks to conduct machine learning experiments, and found that the effect is quite good in visual and speech recognition. After the birth of the BP algorithm, the development of neural networks entered a boom.
2. Principle of Neural Networks
So what is the learning mechanism of neural networks? Simply put, it is decomposition and integration. A complex image into a large number of details into the neurons, neurons processed and then integrated, and finally came to the conclusion that what is seen is correct. This is the mechanism of visual recognition in the brain, and also the mechanism of neural networks work. So you can see that neural networks have obvious advantages.
3. Logical Architecture of Neural Networks
Let's look at the logical architecture of a simple neural network. In this network, it is divided into an input layer, a hidden layer, and an output layer. The input layer is responsible for receiving signals, the hidden layer is responsible for decomposing and processing the data, and the final result is integrated into the output layer. A circle in each layer represents a processing unit, which can be thought of as simulating a neuron, and several processing units form a layer, and several layers form a network, which is also known as a "neural network". In a neural network, each processing unit is in fact a logistic regression model, the logistic regression model receives inputs from the upper layers and transmits the model's predictions as outputs to the next level. Through such a process, neural networks can accomplish very complex nonlinear classification.
4. Neural network applications.
The field of image recognition is a famous application in neural networks, this program is a neural network constructed on the basis of multiple hidden layers. With this program it is possible to recognize a wide range of handwritten numbers and achieve high recognition accuracy with good robustness. It can be seen that as the layers get deeper and deeper, the deeper the layers the lower the detail handled. However, entering the 1990s, the development of neural networks entered a bottleneck. The main reason for this is that despite the acceleration of BP algorithms, the training process of neural networks is still difficult. Therefore support vector machine algorithms took the place of neural networks in the late 90s.
In this article we have introduced the knowledge about neural networks, the specific content is the origin of neural networks, neural networks, the principle of neural networks, neural networks, logical architectures and neural networks, I believe that you see here on the knowledge of neural networks have a certain understanding, I hope that this article can help you.