Two types of learning approaches for data science and machine learning

There are two main methods in mastering new knowledge:

Bottom-up and Top-down learning.

What’s Bottom-up learning

According to Encyclopedia of the Sciences of Learning,

Bottom-up learning refers to learning implicit knowledge first and then learning explicit knowledge on that basis (i.e., through “extracting” implicit knowledge).

Most of our schools are built around the bottom-up natural progression through material. A host of technical and scientific fields of study are taught this way.

I take cancer biology as an example. There is a logical way to lay out the topics in cancer biology that build on each other and lead through a natural progression in skills, capability, and understanding. Most student will gradually find out what’s cancer, why and when cancer happens, how we treat cancer, the cutting-edge fields of the treatments (immunotherapies, etc.) and the future of cancer biology.

“The problem is, the logical progression through the material may not be the best way to learn the material in order to be productive.”

You can learn technical subjects from the bottom-up, and a small percentage of people do prefer things this way, but it is not the only way. At least until you think about how you actually learn.

Let’s recall how we learned to read, drive, cook..

We learned practical skills rather than figured out the definitions of technical terms, right? When we get more interested and confident, we then turn to books for theory and minor details. That’s a top-down manner.

What’s top-down learning?

According to Encyclopedia of the Sciences of Learning,

Top-down learning refers to learning explicit knowledge first and then learning implicit knowledge on that basis (i.e., assimilating explicit knowledge into an implicit form).

I prefer this approach in learning data science and machine learning.

What’s the benefits of top-down learning?

Data science is a fast-pacing field. The techniques you learned a month ago may be substituted by new ones. We are not robots that can memorize all the theories. My suggestion is: Don’t start with definitions and theory. Instead, start by connecting the subject with the results you want and show how to get results immediately.

Lay out a program that focuses on practicing this process of getting results, going deeper into some areas as needed, but always in the context of the result they require. You go straight to the thing you want and start practicing it. You have a context for connecting deeper knowledge and even theory. You can efficiently sift and filter topics based on your goals in the subject.

Practice on many small projects and slowly increase their complexity. I bet it will make you much better.

Leave a Reply

%d bloggers like this: