(Python tutorial) what are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

What is machine learning, what is artificial intelligence and what is deep learning? These terms and concepts have been heard all the time, but they are also easy to be confused

artificial intelligence

Artificial intelligence technology hopes to use computers to construct complex machines with the same essential characteristics as human intelligence. These machines and algorithms that can replace artificial work are uniformly called artificial intelligence. Artificial intelligence is a large category. With the continuous development of computer technology, the research field of artificial intelligence is also expanding, The following figure shows various branches of artificial intelligence research, including expert system, machine learning, evolutionary computing, fuzzy logic, computer vision, natural language processing, recommendation system, etc.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

artificial intelligence

Machine learning: a method to realize artificial intelligence

The most basic approach of machine learning is to use algorithms to analyze data, learn from it, and then make decisions and predict events in the real world. Different from the traditional hard coded software programs to solve specific tasks, machine learning uses a large amount of data to "train" and learn how to complete tasks from the data through various algorithms.

Deep learning, a neural network method to realize machine learning

For example, in the field of computer vision, if a panda is recognized, the method of machine learning is to tell the machine various characteristics of the panda, such as nose, eyes, mouth, hair and so on, so that the machine can realize that it is a panda with these characteristics

However, the method of deep learning is to give the machine a picture, let the machine extract the features, and then predict whether it is a panda. If the prediction fails, the neural network will tell the neural network where there is an error through forward transmission, and re identify until the recognition is correct. The most famous is CNN convolutional neural network, including computer recognition, Natural language processing, expert system, recommendation system and so on all make more or less use of the knowledge of CNN convolutional neural network

The relationship between the three is well explained by summarizing the figure below. With the continuous improvement of computer algorithms, deep learning is more and more appreciated in the field of artificial intelligence.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

The core goal of AI is to provide a set of algorithms and technologies that can be used to solve the problems that human beings can automatically execute intuitively, but are very challenging for computers. A good example of this kind of artificial intelligence problem is to interpret and understand the content of images - a task that human beings can complete effortlessly, but it has proved difficult for machines to complete.

Artificial neural network (ANN) is a kind of machine learning algorithm. It learns from data and focuses on model recognition. Its inspiration comes from the structure and function of the brain. Deep learning is a subset of artificial intelligence. We focus on deep learning.

A concise history of neural networks and deep learning

"Deep learning" has existed since the 1940s and has undergone various name changes, including cybernetics, connectionism and the most familiar artificial neural network (ANN). Although inspired by how the human brain and its neurons interact, artificial neural networks do not mean a realistic model of the brain. Instead, they are an inspiration that allows us to compare a very basic brain model with how we imitate some of these behaviors through artificial neural networks.

The first neural network model came from McCulloch and Pitts in 1943. This network is a binary classifier, which can recognize two different categories according to some inputs. The problem is that the weights used to determine the class labels for a given input need to be manually adjusted - this type of model obviously cannot be well extended if manual intervention is required.

Then, in the 1950s, Rosenblatt (1958, 1962) released a pioneering perceptron algorithm - a model that can automatically learn the weights required to classify inputs without human intervention. An example of the perceptron architecture can be seen in the figure below. In fact, this automatic training process forms the basis of random gradient descent (SGD), which is still used to train very deep neural networks.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

perceptron algorithm

A simple example of a perceptron network architecture that accepts multiple inputs, calculates a weighted sum, and applies a step function to obtain a final prediction.

During this period, perceptron based technology was all the rage in the neural network community. However, a paper published by Minsky and {Papert in 1969 effectively stagnated the research of neural network for nearly a decade. Their work shows that the perceptron with linear activation function (regardless of depth) is only a linear classifier and can not solve the nonlinear problem. A typical example of a nonlinear problem is the XOR data set in the figure below. It is impossible to try a straight line to separate the blue star from the red circle.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

XOR dataset

In addition, the author believes that we do not have the computational resources required to construct large-scale deep neural networks, and this paper alone almost stifles the research of neural networks.

Fortunately, the back propagation algorithm was proposed by werbos, Rumelhart and Lecun. It can make the neural network recover from the situation that may have died prematurely. Their research on back propagation algorithm enables multilayer feedforward neural network to be trained.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

Back propagation algorithm

A multilayer feedforward network architecture has an input layer (3 nodes), two hidden layers (the first layer has 2 nodes and the second layer has 3 nodes) and an output layer (2 nodes).

Combined with nonlinear activation function, researchers can now learn nonlinear function and solve XOR problem, which opens the door to a new research field of neural network. Further research shows that the neural network is a general approximator, which can approximate any continuous function (but does not guarantee whether the network can really learn the parameters required to represent the function).

Back propagation algorithm is the cornerstone of modern neural networks, so that we can effectively train neural networks and "teach" them to learn from mistakes.

Perhaps a typical example of applying deep learning to feature learning is the convolutional neural network (Lecun) applied to handwritten character recognition, which automatically learns the discrimination mode (called "filter") from the image by stacking layers at the top of each image in turn. The lower layer of the network filters represent edges and corners, while the higher layer uses edges and corners to learn more abstract concepts used to distinguish image categories.

While working at Bell Labs, Lecun developed a system that can recognize handwritten digits and named it lenet. Maybe you haven't heard of lenet, but it was used by most American banks to recognize handwritten digits on cheques. To reach this commercial level, its accuracy can be imagined. So what exactly is lenet? Lenet is a typical convolutional neural network used to recognize handwritten digits. It is the first time that convolutional neural network is applied to solve practical problems. The most famous MNIST data set is the data set trained by this neural network

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

MNIST

The development of neural network has successfully opened the door to deep learning. With the vigorous development in recent years, different computer algorithms, especially the continuous development of CNN convolution neural network, have gradually expanded the development of deep learning

Other directions of deep learning, natural language processing

Natural language processing (NLP) is an important direction in the field of computer science and artificial intelligence. It studies various theories and methods that can realize effective communication between human and computer with natural language. Natural language processing is a science integrating linguistics, computer science and mathematics. Therefore, the research in this field will involve natural language, that is, people's daily language, so it is closely related to the research of linguistics, but there are important differences. To put it bluntly, natural language processing is a technology that tells machines how to hear, including the first smart voice assistant in the mobile phone industry, smart speaker, smart home, etc

Other directions of in-depth learning, recommendation system

We have access to the recommendation system in our daily life, especially when we use headlines, jitter, shopping platforms, etc., we tiktok videos, read articles, purchase records, etc., have become the data of the recommender system learning, and then recommend more video articles that meet our expectations.

What are machine learning, artificial intelligence and deep learning, and what is the relationship between them?

Recommendation system

Finally, I would like to recommend my own Python learning group:[eight hundred and fifty-six million eight hundred and thirty-three thousand two hundred and seventy-two], everyone in the group is learning python. If you want to learn or are learning python, you are welcome to join. Everyone is a software development party. They share dry goods from time to time and receive free live courses. Including a copy of the latest Python advanced materials and zero basic teaching in 2021 compiled by myself. Welcome to join us! You can also scan code and VX to get data!


 

 

  • one
    give the thumbs-up
  • 0
    comment
  • three
    Collection
  • One key three links
    One key three links
  • Sweep and share posters

<p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff; text-align: center;"><span style="margin: 0px; padding: 0px; font-size: 14pt;"><strong style="margin: 0px; padding: 0px;"><span style="margin: 0px; padding: 0px; color: #993366; "> author's introduction < / span > < / strong > < / span ></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><span style="margin: 0px; padding: 0px; color: #993366; "> author Toby: licensed consumer finance model expert, maintaining long-term project cooperation with professors of Chinese Academy of Sciences and China University of science and technology; It has project docking with tongdun, juxinli and other external data source companies. Familiar with consumer finance scene business, online and offline business, including cash loan, commodity loan, medical beauty, anti fraud, auto finance, etc. Model project 200 +, good atPythonmachine learningModeling has good solutions to difficult problems such as variable screening, derived variable construction, high missing rate of variables, imbalance of positive and negative samples, high collinearity, multi algorithm comparison, parameter adjustment and so on</ span></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><span style="margin: 0px; padding: 0px; color: #993366; "> the author's gift -- and its recipientartificial intelligenceInstead, it's better to take the initiativestudyProgramming, designmachineServe yourself < / span ></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><span style="margin: 0px; padding: 0px; color: #993366;"> </span></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff; text-align: center;"><span style="margin: 0px; padding: 0px; font-size: 14pt;"><strong style="margin: 0px; padding: 0px;"><span style="margin: 0px; padding: 0px; color: #993366; "> course background < / span > < / strong > < / span ></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff; "> I've been there many timespythOn training, explaining programming to students has a great impact on life. I was confused when I graduated from college. I didn't know what to do in the future. In the face of the dense crowd at the job fair, I was at a loss. Since contactpythAfter on programming, I was surprised by this fast and efficient programming language. Since then, the trajectory of my life has changed quietly. Programming is not the monopoly of computer major. I remind students many times. In Europe, America and Japan, students majoring in art, music, English, archaeology, mathematics and physics also use programming. Programming is just a tool that allows us to quickly implement the logic algorithm of the brain. Don't doubt yourself any more. Go ahead and knock down the first line of code "hel"loworld! Congratulations, you are already a programmer, right, that's it, follow me!</ p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;">pythOn programming let me find a wonderful world once confused. I'm happy to share these valuable resources and experience. I hope to help you who are also confused! In order to make the majority of students free and faststudypythOn, I have prepared a new course for you < a style = "margin: 0px; padding: 0px; Color: #000000;" href=" https://study.163.com/course/courseMain.htm?courseId=1006183019&share=2&shareId=400000000398149 " target="_blank" rel="noopener">《PythOn entry classic (2k ultra clear) < / a >. The course catalogue is as follows, which roughly includespythOn environment construction, resource introduction, basic knowledge and employment guidance</ p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff; "> the video is recorded with a special graphics card and supports 2K ultra clear resolution. Students can see each line of code and text clearly and have a better user experience</ p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"> </p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff; text-align: center;"><span style="margin: 0px; padding: 0px; color: #993366; font size: 14pt; "> < strong style =" margin: 0px; padding: 0px; "> Course Overview < / strong > < / span ></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: "#ffffff;" > this course is free from explanationpythOn grammar, but in another waypython。 Let beginners fully understandpythWhat fields can on be used in and what are the benefits of learning< Br style = "margin: 0px; padding: 0px;" / > this course is suitable for beginners to lay a solid foundation,alsoCan helpPythOn programmers improve their skills, even intermediate and advancedPythOn programmers can also find refreshing content from books< Br style = "margin: 0px; padding: 0px;" / > Introduction to Chapter 1 of the coursepythOn official website, software download address, and special data science advanced version framework anaconda. How to install PIP for beginnerspythOn's third-party package< Br style = "margin: 0px; padding: 0px;" / > Chapter 2 recommends somepythonstudyBooks and free database resources formachine learningandartificial intelligenceModeling. Chapter III introductionpythOn quick look-up table to save time for rookies and old birds to check grammar; And basic grammar, showingPythOn DIY plant vs zombie game. Finally providePythOn employment guidance to provide better golden jobs than civil servants< Br style = "margin: 0px; padding: 0px;" / > the author wants to share all thePythOn knowledge to everyone, but time is limited, finally share somestudyMethods to students, so that everyone can respond to changes with invariance</ p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><img src=" https://img-bss.csdnimg.cn/202010080420245753.png " alt="" width="880" height="1189" /></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"> </p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><img src=" https://img-bss.csdnimg.cn/202010080420462404.png " alt="" width="880" height="2025" /></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><img src=" https://img-bss.csdnimg.cn/202010080420595183.png " alt="" width="880" height="1407" /></p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"> </p> <p style="margin: 10px auto; padding: 0px; color: #333333; font-family: 'PingFang SC', 'Microsoft YaHei', 'Helvetica Neue', Helvetica, Arial, sans-serif; font-size: 14px; background-color: #ffffff;"><img src=" https://img-bss.csdnimg.cn/202010080421126529.png " alt="" width="880" height="992" /></p>
emoticon
Insert expression
©️ 2020 CSDN Skin theme: dark blue ocean Designer: CSDN official blog Return to home page
Paid inelement
Payment with balance
Click retrieve
Code scanning payment
Wallet balance 0

Deduction Description:

1. The balance is the virtual currency of wallet recharge, and the payment amount is deducted according to the ratio of 1:1.
2. The balance cannot be purchased and downloaded directly. You can buy VIP, c-coin package, paid column and courses.

Balance recharge