1. Artificial intelligence
1.1 What is artificial intelligence?
The concept of artificial intelligence was proposed in a seminar held at Dartmount University in the United States in 1956. Artificial Intelligence (AI), referred to as AI, is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a manner similar to human intelligence. Research in this field includes Robotics, speech recognition, image recognition, natural language processing and expert systems, etc. It is mainly used to help humans complete complex calculations, improve work efficiency, and retrieve massive amounts of information .
1.2 Functions of artificial intelligence
- Speech recognition : Human speech can be recognized and understood, and converted into processable text or instructions.
- Natural language processing : It can understand and process natural language, including text analysis, semantic understanding, grammatical error correction, etc.
- Intelligent recommendation : According to the user's personal preferences and behavior patterns, suitable products, services and content can be recommended.
- Automatic driving : It can realize automatic navigation and control of vehicles such as cars and drones, and reduce the occurrence of traffic safety accidents.
- Smart finance : It can analyze large amounts of data, conduct risk assessment, financial analysis and investment decision-making, etc.
1.3 The role of artificial intelligence
1.3.1 Empowering economic development and serving people's lives
As an enabling means, the integration of artificial intelligence and the real economy can lead industrial transformation and breed new industries, new models and new formats. As a tool to serve people's better life, the application of artificial intelligence can help improve the quality of life and meet people's consumption upgrade needs.
1.3.2 Improving the efficiency of government governance
In recent years, with the integration of new-generation digital technologies such as big data, cloud computing, and artificial intelligence into the construction of digital government, "handle management" and "fingertip management" have become standard government services. Flowing data, smooth experience, allowing people to run less errands, and data to run more, are bringing people a real sense of gain.
1.3.3 Promote educational innovation
As an important driving force leading a new round of technological revolution and industrial transformation, artificial intelligence has spawned a large number of new products, new technologies, new industries and brand-new models, and has also brought more possibilities for the modernization of education. With the support of artificial intelligence, there will be more high-quality digital educational resources, promoting education to be more equitable.
1.3.4 Information Misuse
Face and other information are unique, and now that face recognition is widely used, some security problems will arise.
2. Machine Learning
2.1 What is machine learning?
Machine learning is an artificial intelligence technique that uses algorithms to allow computers to learn from data so that decisions and predictions can be made automatically. The purpose of machine learning is to use data to improve the accuracy of predictions or decisions without human intervention.
Machine learning algorithms can be applied in various fields such as natural language processing, image recognition, medical diagnosis, etc. They can automatically extract patterns from large amounts of data, and use these patterns to make predictions and decisions, thereby helping people better understand and apply data. For example, when you use the auto-tagging feature on a social media platform, you may find that the system tags people and objects in the photos you upload. Behind this, machine learning can actually be reflected.
2.2 Machine Learning Algorithms
2.2.1 Linear regression
Linear regression is one of the easiest to understand algorithms in statistics and machine algorithms.
The representation of linear regression is an equation that describes the line of values that best fits the relationship between an input variable (X) and an output variable (Y) by finding specific weights of input quantities called coefficients (B).
2.2.2 Logistic regression
Logistic regression is another technique that machine learning borrows from the field of statistics. It is the method of choice for binary classification problems (problems with two class values).
Logistic regression is similar to linear regression in that the goal is to find coefficient values that weight each input variable. Unlike linear regression, the prediction of the output is transformed using a non-linear function called the logistic function.
2.2.3 Linear Discriminant Analysis
Linear discriminant analysis is a generalization of Fisher's linear discriminant method, which uses statistics , pattern recognition , and machine learning methods to try to find a linear combination of the characteristics of two classes of objects or events in order to be able to characterize or distinguish them . The resulting combination can be used as a linear classifier or, more commonly, for dimensionality reduction for subsequent classification. If you have more than two classes, then the linear discriminant analysis algorithm is the preferred linear classification technique.
2.2.4 Classification and regression trees
The English of classification and regression tree is Classification and regression tree, and the abbreviation is CART. CART is a learning method that outputs a conditional probability distribution of a random variable Y given an input random variable X. CART assumes that the decision tree is a binary tree, the values of the internal node features are "yes" and "no", the left branch is the branch with the value "yes", and the right branch is the branch with the value "no". Such a decision tree is equivalent to recursively bisecting each feature, dividing the input space, that is, the feature space, into a finite number of units, and determining the predicted probability distribution on these units, that is, the conditional probability of the output under the given input conditions distributed. It consists of tree generation and tree pruning. Decision tree is an important algorithm of machine learning.
2.2.5 Naive Bayes
Based on Bayesian theorem, it is assumed that the features are independent of each other and used for classification problems.
2.2.6 Support Vector Machine (SVM)
For classification and regression problems, find an optimal hyperplane for classification by mapping the data into a high-dimensional space.
2.2.7 K nearest neighbors
Classify the target point into the most frequently occurring category by calculating the distance between the target point and the nearest K points in the training dataset.
2.2.8 Random Forest
The integrated learning method predicts through the results of multiple decision trees to improve the classification effect.
2.2.9 Dimensionality reduction
The so-called dimensionality reduction in the field of machine learning refers to the use of a certain mapping method to map data points in the original high-dimensional space to a low-dimensional space. The essence of dimensionality reduction is to learn a mapping function f : x->y, where x is the expression of the original data point, and the vector expression is currently used at most. y is a low-dimensional vector expression after data point mapping, usually the dimension of y is smaller than that of x (of course, it is also possible to increase the dimension). f may be explicit or implicit, linear or nonlinear.
2.2.10 Artificial neural network
Imitating the structure and function of biological neural networks, learning and detection are performed through the connections between multi-layer neurons.
3. Deep Learning
3.1 What is deep learning?
Deep Learning (DL) is a specific form of machine learning that uses neural networks with multiple layers of nonlinear processing units to learn and represent data. At the heart of deep learning is the use of deep neural networks , which can automatically extract complex features from data and perform well on a variety of tasks, such as image recognition, speech recognition, and natural language processing.
3.2 Deep Learning Model
- Convolutional Neural Networks (CNN): Mainly used for image and video-related tasks, extracting spatial features through local connections and weight sharing.
- Recurrent Neural Networks (RNN): Capable of processing sequence data, such as tasks such as natural language processing and speech recognition. RNNs are able to use past information as context to make predictions.
- Long Short-Term Memory (LSTM): A special RNN variant that solves the gradient disappearance and gradient explosion problems of traditional RNNs through a gating mechanism.
- Generative Adversarial Networks (GAN): Composed of a generator and a discriminator, the generator tries to generate realistic samples, while the discriminator tries to distinguish real samples from generated samples, which are optimized by confrontational learning.
- Attention Mechanism : When processing variable-length sequence data, it can give the model the ability to independently select and focus on important information.
Four, the relationship between the three
There is a certain relationship between artificial intelligence, machine learning, and deep learning. Artificial intelligence refers to a discipline and technology in which computers can simulate human intelligence. Machine learning is a branch of artificial intelligence designed to enable computers to automatically learn and improve performance through data and experience, without explicit programming instructions. Deep learning is a special form of machine learning , which learns and makes decisions by simulating the structure and function of the human brain neural network.
Simply understood, machine learning is a method to realize artificial intelligence, and deep learning is a technology or algorithm of machine learning. Deep learning utilizes a large amount of data and a multi-level neural network, which can perform better feature extraction and pattern recognition, and has stronger representation learning and decision-making capabilities than traditional machine learning methods.