# What math should I learn before and while using and applying deep learning?

I want to learn deep learning. After researching a little, I came to the conclusion that I need a lot of math. I've started a linear algebra course, and it takes a long time (2-3 weeks). I want to start using and applying deep learning to solve problems in this summer, but I assume I would not have enough time to learn all subjects (linear algebra, statistics and probability and calculus 1).

So, what math should I learn before and while using and applying deep learning?

• In my opinion the answer to this sort of question is often to do both at once. Assuming you've at least done first year university maths I'd say go for it. Learning for the sake of being able to start learning something else is a surefire way of minimizing engagement and uptake... Get started on DL and it will force you to learn the math required (as long as you're honest with yourself) Jun 28, 2021 at 9:26
• What do you mean by "I want to start participating deep learning"? Do you mean participate in deep learning competitions/challenges? Maybe you can provide an example of such competition (if that's the case).
– nbro
Jun 28, 2021 at 11:22
• I meant while making neural networks or beefore. Jun 28, 2021 at 12:53
• Ok, I've modified your post to make it more understandable then. Make sure my changes are consistent with what you were asking.
– nbro
Jun 28, 2021 at 13:06

The math that you need to be comfortable with most deep learning (DL) topics (such as neural networks, gradient descent are back-propagation) is already mentioned in your post, but I will list the main subjects here too.

• Linear algebra (an entire college-level course is necessary; you can start with Khan Academy videos/lessons and you can pick one of Gilbert Strang's books)
• Calculus (same; Kenneth A. Ross' book is a decent one)
• Numerical analysis/algorithms (you need be aware of numerical algorithms, like gradient descent, and concepts like convergence, round-off errors, etc; in fact, gradient descent is the widely used in DL)
• Probability theory (you need to know what a probability distribution, random variable, etc., are)
• Statistics (you don't need to know everything at the beginning, but the more you know the better)

I didn't use this book when I was studying deep learning, but part 1 of this book covers (at least some of) the most important mathematical prerequisites for deep learning, so you could try to read some of the chapters to understand at what point you are. I don't have a favourite book for the last 3 topics listed above.

Check out also the book Mathematics for Machine Learning. I never read it, but it looks like part 1 has many chapters on most important math topics for ML and so DL.

By the way, I don't think that 3 weeks is a lot. You will definitely need more time to learn the mathematical prerequisites for deep learning, but the exact time depends on your specific background.

• Thank you, I was already taking the Khan Academy course :D. Jun 28, 2021 at 12:54
• @Vexea After a long time, I've come across this post again. If my answer answered your question, you should accept it.
– nbro
Jan 24, 2023 at 15:36