Introduction to the era of intelligent everyone should understand what is the depth of learning – So shdoclc.dll

Introduction to the era of intelligent | everyone should know: what is deep learning? Sohu science and technology selected from the heart of the algorithmia machine to participate in: Wu Ping this paper describes the relationship between the three aspects of artificial intelligence, deep learning and machine learning to clarify the depth of learning and its importance. Traditional machine learning intelligent processing of a certain amount of data, and for deep learning, the more data, the better the performance of the depth of learning technology. In addition, this paper also introduces the framework of deep learning and excellent online courses and books. To understand what deep learning is, we first need to understand the relationship between deep learning and machine learning, neural networks and artificial intelligence. The best way to represent this relationship is to visualize them in concentric circles: the outermost ring is artificial intelligence (using computer reasoning). Inside the machine learning. Artificial neural networks and depth learning at the center. Broadly speaking, deep learning is a more accessible name for artificial neural networks. Deep learning refers to the depth of the network. And an artificial neural network can also be very shallow. Neural networks are inspired by the structure of the cerebral cortex. The most basic level is the perceptron (perceptron), a mathematical representation of neurons. Like the structure in the cerebral cortex, the neural network can have several layers of interconnected sensors. The first layer is the input layer. Each node in the layer passes an input, and then the output of the node is passed as an input to each node in the next layer. There is no connection between nodes in the same layer, and the final output is processed. We call the middle part the hidden layer. These neurons do not have external connections (such as inputs and outputs), and are activated only by nodes on the top layer. Source: Michael A. Nielsen, "Neural Networks Deep Learning and", it is believed that deep learning is learning neural network, and the use of multi layer abstraction to solve the problem of pattern recognition technology. In 1980s, due to the limitations of computational cost and data volume, most neural networks have only one layer. Machine learning is considered as a branch of artificial intelligence, and deep learning is a special kind of machine learning. Machine learning involves computer intelligence, which does not know the answer in advance. On the contrary, the program verifies its attempt by running the training data, and accordingly modifies its method based on the success of the program. Machine learning involves a number of disciplines, from software engineering and computer science to statistical methods and linear algebra. There are two kinds of machine learning methods: in supervised learning, machine learning algorithms use labeled data sets to train rules. This requires a lot of data and time, because the data need to manually mark. Supervised learning is an excellent choice for classification and regression problems. For example, suppose we are running a company and want to understand the impact of bonuses on employee retention. If we have historical data on employee bonus amounts and tenure we can use supervised machine learning. Unsupervised learning does not have any predefined n相关的主题文章: