How to learn and understand mathematics for ML for a ninth grader?

Question

IT education

Now in the ninth grade, mathematics at the level of physics and mathematics of the lyceum is solid 4. I am very interested in the topic of Data Science and Machine Learning. I really like them, but there is a problem in the form of mathematics. I'm pretty good at Python and I know basic SQL queries, but here's the math...

How can I quickly and effectively master the main topics for ML (linear algebra, statistics, math analysis and probability theory)? Any resource/book will help me. English or Russian is not important from the word at all. Unfortunately, I don't have a lot of money to participate in any paid courses.

I understand the complexity of all these topics, but this topic is insanely interesting. I'm ready to go for it!

Answer 1

There are two options:
1. Declare yourself a child prodigy , find your own unique way of studying mathematics - but we are not your advisers here, you yourself somehow
2. Regretfully begin to repeat the thorny path that graduates of the physical and mathematical sciences go through - namely, it’s
normal to learn mathematical analysis in order to understand all these derivatives, integrals, series and other bedlam - and solve physical and geometric applied problems with their help. It is

normal to learn linear algebra so as not to look at the product of matrices as a new gate, but to understand what it is and why it is. Again, solve geometric problems and remember that all these neurons are just a twisted name for the dot product of multidimensional vectors.
Next comes differential geometry, without understanding which it will be difficult to understand optimization methods - namely, why this gradient descent works and what it teaches there.
Combinatorics, statistics, probability theory can be learned thoroughly only if there is a base of matan and linal.
As you can see, the disciplines I have listed go exactly in the order in which they are studied by students of any decent physics and mathematics (those who got there after 11 classes) - simply because this is a fucking logical order of study, based on previously acquired knowledge.
In total, for a normal (conscious) orientation in ML, you need to have the knowledge of a 2nd year student of any physics and mathematics (even a city pedagogical institute is enough for you). If you want a similar result in the ninth grade (and most importantly, fast!!!!), you should have started in the fifth.

Answer 2

Easiest. Dmitry Pismenny - textbook + problem books, read a book, solve problems. Master the Written and then just stupidly learn tesorflow. Your chance of getting a job increases the more your knowledge of libraries increases - sklearn/tensorflow/torch etc. The main thing here is not to spend too much time on mathematics. Yes, it is needed, but let's be honest - by and large, everyone doesn't care if you can answer by heart how back propagation works there. As proof, you can go to hh and check vacancies - it is important that you are able to solve practical problems there.
ps
Well, English! Mandatory. Already from the answers to this question, you already probably understood that NO ONE IS GOING TO ANSWER you in Russian, they will try to humiliate you at most and show what a fool you are.

Answer 3

How can I quickly and effectively master the main topics for ML (linear algebra, statistics, math analysis and probability theory)?

Finish school, having mastered elementary mathematics and enter the Mekhmat of Moscow State University. There they will teach this to everything (if you study, and not play the fool). It's pretty fast and efficient.

Answer 4

>Now in the ninth grade, mathematics at the level of the physics and mathematics lyceum is solid 4
if you do not need to invent new algorithms, or write some of your own libraries, then your knowledge of mathematics is already enough to work in ML topics. You can focus on the python and its ecosystem, all sorts of ML libraries. Just sort out specific tasks, repeat after others.

Answer 5

In general, look to find out what a series is, operations on a series, addition, the sum of a series, addition / multiplication / difference / division of a series by another series or by itself. Like equations, where not a single number is known, but a series. Learn the limits, then the matrix, operations on it as in series. A bit of logic, boolean operations. Types of neural networks, then file formats (sound, images, video) and a programming language like python or matlab). Documents excel, word, pdf, csv (how to extract data, how to write this also applies to the file format). SQL language, English language, framework (library) or neural network module. Study the subject area you are interested in and come up with a model of neural networks that will solve problems for it