Thread by @svpino

The one million dollar question:

"Is it reasonable for someone to dive into machine learning with a shallow knowledge of math?"

▫️ The short answer is "yes."
▫️ The more nuanced answer is "it depends."

Let me try and unpack this question for you.

🧵👇

You can think about machine learning as a spectrum that goes all the way from pure research to engineering.

The more you move towards a research position, the more you can benefit from your math knowledge. If you move in the other direction, you'll get away with less of it.

👇

I have friends that got a Ph.D. and became college professors.

For them, math is an absolute requirement!

Not only are they working on research projects, but they are teaching the next generation of scientists and engineers.

👇

Other friends went the other direction: they took positions at companies that focus on exploiting existing high-level frameworks to produce value.

A lot of the math required is already abstracted away by these libraries. They can get away with much less knowledge.

👇

Notice that I didn't say that you don't need math *at all*. I specifically framed the initial question as having a "shallow knowledge" of it.

It's tough to escape from math, regardless of what you do in life. Machine learning is no exception.

👇

I'm an engineer. This is my experience:

▫️ Basic knowledge of statistics and probabilities is essential for what I do.

▫️ Understanding derivatives much less so.

▫️ Linear algebra comes up all the time, but it's easy to refresh concepts when I need them.

👇

I want you to keep this in mind:

▫️ You can get started with a much shallower knowledge of math than what would be required for a job.

Part of the process is to learn what you need!

This is probably the most relevant piece of advice that always gets lost in translation.

👇

If today is your first day getting on the machine learning train, I promise you can get away with basic high school level math.

In a year, you'll need more than that. But that will happen gradually.

👇

Remember when you were learning to ride a bicycle?

If you look back and try to think logically about it, there are so many things you probably *wish* you knew before starting.

But you didn't.
And you learned.
And everything turned out okay.

👇

Let's wrap up this with a TLDR;

▫️ You can get started with very little math.
▫️ As you progress, you'll incorporate more of it.
▫️ Your specialization will determine how much.

More important than anything:

GET. STARTED. Don't overthink it.

I finish work every day around 5 pm. Then I come here to share some of the things I've learned.

Every day ☀️🌙

If you don't mind a blunt and practical point of view about machine learning and software engineering, stay tuned!