Textbooks, tutorials, Chinese, and a wave of direct answers to questions:
Note: I don’t think there are many tutorials, but they are refined. There are many tutorials on the Internet, but below I will organize a Chinese machine learning and deep learning basic tutorial that I think is more reliable. While learning the basics, you should learn English well!
Start with python zero foundation, recommend Liao Xuefeng python tutorial, 2, 3 optional
1 Machine Learning:
"Machine Learning Practical Combat" is a must for getting started, not difficult, easy to use, and can increase interest. Disadvantage: the theory is not deep enough
The "Machine Learning" Watermelon Book, a classic in theory of "Statistical Learning Methods", with detailed and in-depth theory, is really worth seeing. Disadvantages: A bit boring, it is recommended to combine machine learning with actual combat.
2 Deep learning frameworks and libraries (pick up the code):
First come to the code, I always feel that if the theory is boring, then use the code to raise interest first.
TensorFlow official document Chinese version, Official, reliable! tensorflow artifact, absolutely must have!
Machine Learning Series | Mo Fan Python, I recommend Mo Fan tutorial again. There are many tutorials here, including numpy and pytorch tutorials. Advantages: simple, easy to understand, recommend pytorch to learn, more flexible than tensorflow, personal feeling. Disadvantages: Some are too simple.
3 deep learning:
Wu Enda's deeplearning.ai, which has Chinese subtitles in NetEase Cloud Classroom-- link. This introductory deep learning can be there. Teacher Wu’s class is popular, easy to understand, and most importantly: reliable
Classic cs231n, official website link, Cs231n has notes- link, Of course, some great gods on Zhihu translated this note--- link . If you have trouble reading English, refer to this translation, the Chinese version.
"Deep Learning" is a very good theoretical book for deep learning, and github has the original translationIt is the same as what is sold in the market now, which is very good.
What word2vec, CNN, RNN, etc., as long as you learn it well, don't have to look at other messy things, a lot of misunderstandings. I feel again: learning is not more refined
4 Reinforcement learning:
Code: Reinforcement Learning Tutorial Series, Mo Fan tutorial, simple
Theory: Recommend two columns and classes:
CS 294 Deep Reinforcement Learning Chinese Notes
David Silver intensive learning open class Chinese explanation and practice
Slowly learn these two, absolutely enough, remember, learn well.
It was enough to finish the above, but there is also at station B, Li Hongyi Deep Learning (2017)_Lecture•Open Class, I’ve seen some of them are not bad, but I think the above is more systematic, and there are homework. I think the combination of practice and theory can learn well.
I don’t agree with some people who say that they don’t know English, so don’t learn it. Whose English is buffed from the mother's womb and will be born? If you don't know how to learn, who said that deep learning is about deep learning from the beginning? Starting from the basics, from Xiaobai to Da Niu, step by step!
I don’t agree to learn deep learning directly without learning English, and I can’t go deep. English and mathematics are very important for machine learning . I don’t know how to learn English, learn English, start by memorizing words, a bunch of methodologies, and search by myself. If you study a university level 4 or 6 and learn some terms and nouns, there is no problem in reading most papers!
If the subject of the question decides to follow the path of machine learning or deep learning, start from the basics, learn mathematics, learn English, write code, find a sense of accomplishment, learn the basics, learn English at the same time , and learn more If the time is long, take it slowly, but if the time is short and you want to make it fast in 5 days and 8 days, then think about it, because it takes one or two months to enter the door. To be honest, the threshold for machine learning deep learning is indeed a bit high. If you don't want to pay any price, just say how powerful it is. That's a dream, wake up early, but if you lose your mind, what about "the abyss"? Besides, it is not to say that there are any great difficulties! Still very interesting!
You can add other knowledge when you have time, so let’s study first ^_^
Recommended reading:
Selected dry goods|Summary of dry goods catalog in the past six months
欢迎关注公众号学习交流~
欢迎加入交流群交流学习~
