Cognitive Science Post 3: Using elaboration to reinforce your understanding of concepts

This is the third and final post on how you can use principles from the cognitive sciences to enhance your learning of the AWS Cloud.

Earlier in this series, we looked at how important it is to not rely solely on passively taking in information from a presentation or a class. Instead, challenge yourself to retrieve (or recall) that information to strengthen long-term learning. We built on that idea by discussing how you can be more efficient and effective with your learning by spacing out your retrieval practice.

Both of these strategies emphasize the important role memory plays in the learning process. To excel in a domain, you must start with a strong foundation of key facts and concepts related to that domain. Take machine learning (ML) as an example. You can’t feature engineer your ML model if you don’t know what feature engineering is in the first place.

But with that said, not all methods for remembering key information through retrieval practice are equal. There are some techniques that can help foster a deeper understanding of information, one of which is known as elaboration.

What is elaboration?

Elaboration is the process of adding detail to the information you’re learning by drawing connections between that new information and existing knowledge. With elaboration, the emphasis is more often on the how and why behind the topic you’re learning and less on the what. Let’s walk through a quick example to illustrate this concept.

Elaboration in action

If you’re new to machine learning, you’ll quickly realize that one of the first hurdles is understanding the domain’s lexicon of terms and concepts. So, your study plan might include reviewing Demystifying AI/ML/DL or What is Machine Learning?, taking note of the key terms and concepts therein, and then quizzing yourself at spaced intervals on those key terms and concepts.

One of your quiz questions, let’s say, reinforcing and strengthening your understanding of the different types of machine learning, might look like this (answer option “a” is the correct answer):

Which of the following is most suitable for supervised learning?

1. Identifying birds in an image
2. Grouping people into smaller groups based on buying habits
3. Reducing the number of features in a data set
4. Identifying anomalies in your data to label credit card transactions as fraudulent

It’s very possible to get this and other similar questions correct, yet not truly have a very deep understanding of what constitutes supervised learning. Elaborate on this topic by challenging yourself to answer follow-up questions, thereby creating a deeper level of understanding.

Here are some follow-up elaboration questions you could use in this or other similar situations:

1. Explain how “Identifying birds in an image” is a good example of supervised learning?
2. Why are the other answer options not suitable for supervised learning?
3. In addition to the correct answer, what’s another suitable use case for supervised learning?
4. Why is the correct answer not an example of unsupervised learning?

How elaboration impacts our brain

The reason elaboration questions are so impactful is due to the way our brains store and retrieve information most effectively. Information that is more densely connected to other information (picture a large and tightly connected web of neurons) is much more easily retrieved in the future than if that information is stored in a shallower network, lacking a rich web of connections. Elaboration questions help you form that dense connection of neurons by challenging you to provide more detail around the topic you’re learning.

So, how do you leverage this principle of elaboration for your own learning of the AWS Cloud? Here are some ideas:

1. Challenge yourself with follow up questions. When learning through retrieval practice – whether you’re answering quiz questions, testing yourself with notecards, or working on a challenging hands-on lab – stop to elaborate on the information you’re retrieving. Don’t just answer the question or complete the lab task, but ask yourself follow-up why and how questions. Challenge yourself to give concrete examples, and connect that new information to what you already know.
2. Draw visual models. One way of elaborating on, and therefore adding rich detail to the information you’re learning, is to generate visual imaginary of that information. For instance, if you’re taking Architecting on AWS, and you’re trying to wrap your head around how a basic three-tier web application works within a VPC that includes subnets, security groups, route tables, and other related services and features, sketch it out! Incorporating drawing into your retrieval practice is a powerful way of enhancing long-term memory of information.
3. Choose project-based instruction. Choose learning opportunities that are student-centered and allow you to not only take in information but give you opportunities to explore and generate your own knowledge with guided instruction. A project-based course like The Machine Learning Pipeline on AWS affords you a lot of opportunity to be an active participant in your learning. It’s these types of environments that make it easier to elaborate on the information you’re learning.

I started this series of blog posts off by admitting how hard it can be to keep up with the hundreds of AWS services and features that are constantly being released. You’re no doubt reading the AWS Training and Certification Blog because you’re excited to begin, or continue, your AWS Cloud learning. The skills and knowledge you’ll acquire are in high demand. My hope is that the cognitive science tips I’ve shared in these three posts will help you pursue your learning with new confidence and enthusiasm, and the know-how to foster long-term understanding to be an AWS builder.