Explorative Learning for Professional Growth

Blog Team

By Vladimir Myltsev

‍

Turn learning into an adventure! Frame your learning journey as an exciting exploration mission, making the process more engaging and enjoyable.
Master a top-down approach. Get the big picture first by understanding the high-level concepts before diving deeper into the details.
Question everything! Develop a curious mindset that drives your learning. Ask those "why" and "how" questions to discover the essence of a topic.
Mine knowledge is like a pro. Use AI tools to research your questions, turning the internet into a treasure trove of valuable information.
Power up your memory. Learn about spaced repetition and use tools like the Recall plugin in Obsidian to truly embed the knowledge you gain.

‍

Picture yourself as an explorer, venturing through space. You’ve just arrived at the mysterious asteroid cataloged as MA-M8766. You’re stuck because your warp drive requires a very rare resource - blue crystals! However, you’ve run out of these crystals and must find a solution to this crisis. Fortunately, you have a friendly AI companion with access to humanity’s vast knowledge base. You’re not alone in space, but it’s time to get down to business!This story describes our common struggle with learning new things, exploring new areas, and mastering new skills. In this article, I’ll share my personal learning system. Remember, the best learning system is the one that works for you. However, I hope you find my approach useful. Learning should be enjoyable and efficient, so avoid overcomplicating it. By the way, I’m currently preparing for the AWS Machine Learning Exam - hence my asteroid code is AWS-MLS-C01. What’s the current asteroid that you’re stuck on?

‍

First of all, let’s take a look at our ship. It looks amazing, we have a giant store for crystals, advanced crystal processing technologies, and a lot more interesting features. It seems we can explore our asteroid without worrying about our spaceship’s capabilities! I use Obsidian for organizing my learning process, although I sometimes switch between it and Notion. Each major subject or learning objective gets its own vault. I consider Obsidian as an extension of my brain, not just another productivity app. Let’s see what we have in our setup!

‍

1. Theme - Choose whatever you like! After all, it’s your spaceship!

2. Excalidraw plugin - This is useful for drawing directly in the document.

3. Highlight - A small but useful plugin for highlighting words and phrases.

4. Recall - This indispensable plugin helps me memorize facts using the Anki app algorithm or other spaced repetition algorithms.

That’s it! Now, let’s descend to the asteroid and start exploring.

By the way, have you met our bot? His name is Gemini, much like Google’s LLM. He has access to a vast knowledge database and can even perform additional checks on the data. He’s quite intelligent, and we’re glad to be ‘stuck’ on this asteroid with him!

Let’s delve into the surface of what might seem like a ‘useless’ stone!

First, create a folder named /notes. Now, it’s time to explore the subject. It’s important to start with a top-down approach. This gives you a broader context. You’re not ready to learn complex concepts yet, but you can start by understanding the smaller components of big ideas. Once you master these smaller ones, you’ll be ready to grasp the larger and more important ideas. That’s why your goal at this stage is to get a broad overview.

As I gear up for the AWS MLS (Machine Learning Specialty) exam, I find myself on an exciting learning adventure. But what exactly am I doing here? Well, I’m exploring!

Here’s how it goes: I jot down notes after attending course lectures, binge-watching YouTube videos, and devouring articles and books. Think of it as my high-level exploration overview—a compass that guides me through the vast landscape of topics necessary to conquer this challenging exam.

At this stage, you’ll be listening to lectures, reading books, and even going through exam questions if they exist. Store all your notes in the /notes folder.

Next, review your notes, articles, and exam questions with a focus on asking questions. Approach your notes, questions, articles, etc., with genuine curiosity, and place any questions that arise in the `/questions` folder. You can place questions in one file, or split them between several files. It’s up to you. The main goal here is to generate as many questions as possible. What is A? Why is B? What is the difference between A and B? Why is C called that way? And so on.

For instance, I have two files in my ‘questions’ folder: ‘Common Questions’ and ‘AWS Questions’. I categorize them into these two groups because my examination concentrates on two primary areas: Machine Learning & Artificial Intelligence (ML&AI) and AWS services. These AWS services are utilized throughout the complete ML process, from data ingestion to model deployment and monitoring.

‍

Now that you have a of the crystal deposits, time to explore. You to the first deposit and together with Gemini, mining raw, crystals.

‍

It’s exciting to how many crystals asteroid contains.

Now, it’s time to use Gemini (or any other AI, book, or search engine) to answer your questions. Try using Gemini Advanced to answer your questions. I recommend Gemini (especially the Advanced paid plan) as it strives to provide better explanations than Chat GPT or Bing Copilot by default. Copilot will even run a search to answer your questions. While this can be useful in other contexts, I’ve found it to be less effective in practice because the reasoning process can be blurred by poor-quality information from the internet. But with Gemini, you get the information first, and then you can click on the ‘check’ button to double-check if the generated answer matches a Google search. Quite useful! Gemini (or another LLM model) here is just a tool, like an advanced search engine. You are the one who will extract these ‘crystals’ from the internet. Ultimately, you must decide what to store in your storage and what to commit to your memory.

Claude3 was released while I was editing this article, so give it a try as well! It’s quite smart.

You will find three types of crystals in this step: entities, concepts, and notions. Let’s create associated folders.

<code>Entity</code> - any object (abstract or concrete) that can be unambiguously defined. In my case, RDBMS, Data Warehouse, Data Pipeline, and AWS Glue are examples of entities.

</code>Concept</code> - an abstract, unambiguously defined idea or principle, often just a property of an entity or group of entities. In my case, correlation (in statistics), Normal Distribution, and Machine Learning are examples of concepts.

</code>Notion</code> - a somewhat complex but important component of your knowledge. It’s something that you were able to formulate based on research. It’s probably subjective and even not accurate in some sense, but it’s still an important part of your understanding of the subject. In my case, the difference between RDMS and Data Warehouse is an example of a notion.Wow, you’re doing an excellent job so far! You’ve found and collected a lot of raw crystals on the asteroid. Now, it’s time to load them into your magic device called Recall and process the raw material into crystals for the engine. It’s a lengthy process that requires discipline and patience from you. But the results are well worth the effort.

At this point, you understand the broader context and even have a lot of interesting facts (entities, concepts, and notions) in your storage (Obsidian files). But you need to memorize them. We will be using the Recall plugin for this purpose.

Create the last folder in your Obsidian and name it cards. We will use a feature of our brains, along with this plugin, to process all this new material and memorize it deeply. This feature is known as ‘spaced repetition’ in scientific literature. There are numerous articles, books, and research papers on the internet about this topic. It’s worth mentioning that our brain memorizes information better when we take breaks between recall sessions.

Review your entities, concepts, and notions again. Pick key memoizable pieces and create a card for each piece of information. It’s important to extract those key terms manually as this process enhances your learning.

But you can use LLMs to extract cards in the desired format for you.When you’re in the card’s document:

Open the command palette (Ctrl + P) and type ‘track’.
Choose ‘Recall: Track Note’.
Once you’ve prepared all the cards, open the command palette again, type ‘review’, and choose ‘Recall: Review’.
Recall what the plugin is asking for. Repeat this process daily. The plugin will schedule sessions for each card based on your answers.

Repeat the whole process (from taking notes to questions to cards) as many times as you feel you need to. At some point, you’ll find yourself much more confident, understanding things that you couldn’t even read a few weeks ago.

‍

After 4 weeks of learning that way, I realized that I understand what our Data Engineering team is talking about on our daily calls 😂 And now I’m able to answer 60-70% of the test questions because I understand the broader context as well as some low-level concepts.