To understand the inner workings of neural networks, a fair amount of mathematical concepts is required. Backpropagation alone is a challenging technique if you are not fluent in calculating local gradients. And that's just the start of the journey.

But the more I study neural networks, the more I get the impression that all those difficult mathematical concepts are only required if you are doing actual research in neural networks or want to know what's happening under the hood. If you "just" want to implement an AI utilizing a neural networks, there are several high level programming frameworks and libraries readily available including model zoos for state of the art neural networks (e.g. VGG, GoogLeNet and ResNet), that can be used.

So my question is, does a developer require a deep understanding of all the details nowadays, or have we reached a level, where frameworks take care of those details for us?


It depends on what exactly you want to be. You don't need to be mathematician if you just want to run neural networks. Most data scientists don't understand the mathematics, but they know how to run machine learning frameworks. Generally, only PhDs understand the mathematics.

In the industry, most job positions in machine learning (e.g software engineer, big data engineer etc) don't require advanced mathematics. But you should be very comfortable in programming, and understand simple things like matrix manipulation. Data engineers in the industry generally have good programming skills, but they can't read mathematical equations (or just simple equations).

However, quantitative positions in machine learning do require significant mathematics. That happens both in research and industry.

Good news --- if you don't have a PhD (e.g statistics) nobody would expect you understand the mathematics.


  • If your job is more like a development position, mathematics is an advantage but not absolutely necessary
  • If your job is research, you do need advanced mathematics understanding (e.g. PhD)
  • If your job is not research but still quantitative based. Mathematics is very important.

PS: I work in research. Everybody I know who works in machine learning has a PhD.

  • $\begingroup$ Great answer! This is also useful from the standpoint of a product manager (i.e. know as much as possible about every facet of the product and business, including the mathematical concepts where relevant b/c you may be called upon to drive product direction. But at the end of the day, the product manager's role to merely to act as a conduit, facilitating the people doing the actual work. This often includes acting as a "translator" between the engineering and sales/marketing sides of the house.) $\endgroup$
    – DukeZhou
    Sep 8 '17 at 17:11

Backpropogation and step functions (as well as a basic understanding of human brain neurons) are absolutelly useful, in my opinion. Rusel-Norvig have a great chapter about them in their "Intro to AI" book that will teach the basics. You don't need a PhD.

  • $\begingroup$ Backpropogation certainly doesn't require PhD, but it's just very basics in neural network. NN is way more complicated than that. OP was asking for "...all the details...". The details can be very technical. $\endgroup$
    – SmallChess
    Sep 6 '17 at 7:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.