MATH of ML

AI Discussions
1

Hello everyone,
I’ve just started the Coursera course Mathematics for Machine Learning.
I think it’s a key step for building a strong foundation in ML, since the math really helps to understand algorithms more deeply.

If anyone here has already taken this course, I’d love to hear your experience and advice! :raising_hands:

2

Many people have taken this course.

Personally, I did not find it extremely useful. Many of the concepts taught have nothing to do with machine learning. They are certainly interesting mathematics, but not really applicable here.

  • You will never use loaded dice in machine learning.
  • You will never use Newton’s Method in machine learning.

Also, many of the concepts that are taught use non-conventional notation or terminology compared to the rest of the DL.AI curriculum. So be aware once you complete this course, you will have to re-learn a number of concepts.

Also be aware that the lectures contain a lot of uncorrected errors.

3

I am very good at Mathematics and I took lots of AI courses.

And I think these are the good amount of math you need to know at least.

  • Calculus: Derivative, Gradient, Function Optimization
  • Probability: Bayes theorem, Law of large numbers
  • Linear algebra: Matrix algebra

And you can learn the other math while you take specific AI courses if you needed.

4

I took Mathematics for Machine Learning and Data Science and found it surprisingly valuable—not because every topic directly applies to machine learning algorithms, but because it builds the kind of mathematical intuition that proves useful down the line. You’re unlikely to encounter trick dice or Newton’s Method in day-to-day ML work, but exploring those ideas helps you understand how randomness and optimization really function under the hood.

5

Thank you very much.. I’m still in the first week of my linear algebra course

But where can I learn other concepts?? Do you have the resources??

Thank you for warning about incorrect errors