I finally finished watching the introductory video Machine Learning - Andrew Ng

  It has been 37 days since I told my tutor that I want to learn artificial intelligence. The tutor was very enthusiastic and expressed his willingness to provide some suggestions. He asked me to study Ng's Machine Learning course online first. It has been 37 days. The overall content is difficult, but It can be considered reluctant to look down and finish it. I intend to write an essay, which can be regarded as a summary of these days.

  At the beginning, I went to Zhihu and searched for the evaluation of this course. Most of the evaluations were: "Introductory video", "Get started after watching it, but it's just an introduction", "The content is simple and easy to understand", I saw these evaluations I was relieved (later I found out how big the gap between me and Zhihu God is, I think it is very difficult in many places T^T), and then I found this course in Netease Cloud Class.

  Wu Enda's machine learning course is divided into eighteen chapters, each chapter is about one and a half hours long, and there are also short ones about half an hour. My first thought was: Huh? Only 18 chapters? If you read one chapter a day, you should finish it in 18 days. Later, I found out that I was still too young, because in the process of watching, there are many things that I don’t understand. Sometimes I need to stop to think and take notes. A ten-minute video sometimes takes half an hour to watch. After reading it, and after reading it, I was still at a loss, and at one point I saw a big head.

  At the beginning, Ng spent a chapter introducing a concept about this course, and then spent two chapters introducing the knowledge of calculus in advanced numbers, linear regression, and some matrix knowledge in linear algebra. These two chapters feel It's still very basic, except for some expressions that are not the same as what I learned, others are similar. But I found a problem, just listening like this, going in the left ear and going out of the right ear is definitely useless, so I came to the blog park to register an account and planned to apply for a blog, but I was rejected. . . Then I tried several times, but all of them were rejected, (I just started my blog this evening). Well, if I don't let me blog, I use my hands, paper, and pen. Good memory is not as bad as bad writing, right?

 

  In this way, I began to memorize while listening, and memorized dozens of pages. Although sometimes I don't know what I'm memorizing, it's better to focus while I'm listening. Otherwise, sometimes I'll get distracted when I'm listening, and I'm helpless to go back and read.

  From Chapter 4 onwards, it seems that we are finally stepping into the core of this course, starting to introduce eigenvalues, gradient descent algorithms, learning rates and so on. Chapter 5 began to introduce OCtave, and I realized that this course also requires programming. . . However, Octave seems to be very similar to Matlab, a programming language that is a bit mathematical. Chapter 6 starts to introduce logistic regression, which is different from linear regression algorithm. Later, in order to solve the problem of overfitting, Chapter 7 starts to introduce regularization. In the following chapters, it becomes more and more difficult to introduce neural networks and vector machines. To be honest, from the beginning of these chapters, I have been at a loss. Many things can only be understood in general, and it is difficult for me to understand some details. Unsupervised learning is introduced from chapter thirteen. I think this is the real artificial intelligence. It reduces the human factor and allows the machine to choose more, such as finding clusters or something, and then introduces the method used when there are too many features. Dimensionality reduction and anomaly detection. The sixteenth chapter introduces the recommendation system. At this time, I realized why every time I open something on Baidu, the next time I open the browser, it will recommend relevant things to me. It turns out that these are the results of machine learning, and artificial intelligence has long been infiltrate life.

  In fact, after reflecting on a total of 18 chapters, I watched it for 30 days before I finished it. Except that sometimes I would pause the video to think about it or take notes when I couldn’t understand it. In fact, some days I only read a little or even I didn't see it. The reason is that there is something else to do. In addition to the usual courses that the school requires me to study, as a monitor sometimes I have to organize class materials. Like the last Qingming holiday, I thought I would have a lot of time to watch this video, but the fact is that I only watched one video during the three days of Qingming. Chapter, because I need to prepare for the class style exhibition after Qingming. And for about ten days after the Qingming Festival, I could only watch one or two videos every day, because that time was the preliminary round of the school debate competition, and there were not enough people in the team to play the game. As the vice captain, I definitely have to play. So during that time almost every night I spent my time on the debate team, discussing debate topics, playing mock competitions, and revising arguments. No way, the Debating Team of the School of Mathematics is the love of my life. Therefore, during that time, it is difficult to take too much time to watch videos while maintaining daily learning.

  To sum up, in general, I also stumbled through the 18 chapters of the introductory video. Although I really don't understand many details, I probably have an understanding.

  I feel a lot of emotion. I feel that it is a frontier field after all. It is certain that it will be difficult. There is still a long way to go in the future. Come on!

Guess you like

Origin http://43.154.161.224:23101/article/api/json?id=325104820&siteId=291194637