Basic concept of Artificial intelligence _ deep learning

1. Background

Artificial intelligence involves very extensive content, from mathematics to computer science, there are a lot of basic knowledge need to reserve, before going to read some artificial intelligence books, always feel more laborious, here will be some of the basic knowledge of artificial intelligence to do a summary comb.

2. Basic Concepts

1) Artificial Intelligence

Artificial intelligence is a branch of computer science that attempts to understand the nature of intelligence and produces a new intelligent machine capable of responding in a similar way as human intelligence, which includes robotics, language recognition, image recognition, natural language processing, and expert systems. One of the main goals of artificial intelligence research is to enable machines to perform complex tasks that usually require human intelligence, such as speech recognition, image recognition, and even chess and Weiqi. AI recently developed very quickly in recent years, Google's artificial intelligence robot Alphago easily defeated the human nine-stage Weiqi master, and Google open up its artificial intelligence platform. such as FACKBOOK,IBM technology companies have vigorously developed the field of artificial intelligence in automatic driving, image recognition, speech recognition and other fields.

2) Artificial Neural network

The realization of artificial intelligence, a large part is based on artificial neural network. Artificial neural Network (Artificial Neural Networks, abbreviated as Anns) is also referred to as Neural Network (NNS), which simulates the behavioral characteristics of animal neural network and carries out a distributed and parallel information processing algorithm mathematical model. This kind of network relies on the complexity of the system, by adjusting the connection between the large number of nodes, so as to deal with the information. The artificial neural network system appeared after the 1940s. It is composed of a large number of neurons adjustable connection weights, with large-scale parallel processing, distributed information storage, good self-organization self-learning capabilities. Error back propagation algorithm is a supervised learning algorithm in artificial neural network. The BP neural network algorithm can approximate any function theoretically, and the basic structure is composed of nonlinear change element, which has strong non-linear mapping ability. Moreover, the parameters such as the middle layer, the number of processing units and the learning coefficients of the network can be set according to the specific conditions, and the flexibility is very great, and it has a wide application prospect in many fields such as optimization, signal processing and pattern recognition, intelligent control and fault diagnosis.

3) BP neural network

BP (back propagation) neural network is one of the most widely used neural network models, which was proposed by the team of scientists, led by Rumelhart and McCelland in 1986, as a multilayer Feedforward network trained by the error inverse propagation algorithm. The BP network can learn and store a large number of input-output pattern mapping relationships without revealing the mathematical equations that describe the mapping relationship beforehand. Its learning rule is to use the gradient descent method to adjust the weights and thresholds of the network continuously through the reverse propagation, so that the error of the network is squared and minimized. The BP neural network model topology includes input layer (inputs), hidden layer (hidden layer) and output layer (outputs layer).

4) Gradient Descent method

The gradient descent method is an optimization algorithm, which is often called the steepest descent method. The steepest descent method is one of the simplest and oldest methods to solve unconstrained optimization problems, although it is no longer practical, but many effective algorithms have been improved and modified based on it. The steepest descent method is to use the direction of the negative gradient as the search direction, the steepest descent method is closer to the target value, the smaller the step, the slower the advance.

The gradient of a function is the rate at which a function changes, where the negative gradient direction search is toward a more and less gradient direction, which involves some basic knowledge of higher data, here is a blog introduced more detailed:

http://deepfuture.iteye.com/blog/1593259

5) Convolution neural network

Convolution neural network is a kind of artificial neural network, which has become a hotspot in the field of speech analysis and image recognition. Its weight sharing network structure makes it more similar to the biological neural network, reduces the complexity of the network model and reduces the number of weights. This advantage is more obvious when the input of the network is multidimensional image, so that the image can be used as the input of the network directly, avoiding the complex feature extraction and the data reconstruction process in the traditional recognition algorithm. Convolution network is a kind of multilayer perceptron specially designed for the recognition of two-dimensional shapes, which is highly invariant to translational, proportional scaling, skew or coplanar deformations.

Convolution neural network is a special deep neural network model, its particularity is embodied in two aspects, on the one hand, the connection between its neurons is not fully connected, on the other hand, the weights of the connections between some neurons in the same layer are shared (that is, the same). Its incomplete connection and weight sharing network structure make it more similar to the biological neural network, which reduces the complexity of the network model (which is very important for the deep structure that is difficult to learn), and reduces the number of weights.

Think back to the BP neural network. Each layer of BP network node is a linear one-dimensional permutation state, the layer and layer of network nodes are all connected. Imagine that if the node connection between the middle and the layer of the BP network is no longer a full connection, it is partially connected. In this way, it is one of the simplest one-dimensional convolution networks. If we extend this idea to two-dimensional, this is the convolution neural network that we see in most references.

6) Machine Learning

Machine learning (Machine Learning, ML) is a multidisciplinary interdisciplinary, involving a number of disciplines such as probability theory, statistics, approximation theory, convex analysis and algorithmic complexity theory. Specialized in studying how computers simulate or implement human learning behavior in order to acquire new knowledge or skills, and to rearrange existing knowledge structures to continuously improve their performance. It is the core of artificial intelligence, is to make the computer has the basic way of intelligence, its application in all fields of artificial intelligence, it mainly uses induction, synthesis rather than deduction.

7) Deep Learning

Deep Learning (English: Deep learning) is a branch of machine learning that is based on a series of algorithms that attempt to abstract data at a high level using complex structures or multiple processing layers composed of multiple nonlinear transformations. Deep learning is a kind of the method of representation learning in machine learning (English: Learning representation). A observed value, such as an image, can be expressed in a variety of ways, such as a vector for each pixel's intensity value, or a more abstract representation of a range of edges, specific shapes, and so on. It is easier to learn a task from an instance (for example, face recognition or facial expression recognition) using some specific representation method. One of the benefits of deep learning is the use of unsupervised or unsupervised (English: semi-supervised learning) Feature Learning (English: Feature learning) and efficient algorithms for layered feature extraction to replace manual acquisition features (English: Feature ( Machine learning)).

3. Summary

Related study materials:

1 Deep learning notes and finishing

http://blog.csdn.net/zouxy09/article/details/8775360

2) Deep Learning

http://blog.csdn.net/abcjennifer/article/details/7826917

3 Open Source Machine learning Library

Http://www.open-open.com/lib/view/open1364432241437.html