The Ultimate Guide to Learning About Artificial Intelligence

A quick note: please help me share this resource!

  • The current growth rate of AI means that it is getting exponentially more valuable as time goes on. This means that investing time into it now will pay dividends later. The sooner we can learn about this impactful technology, the sooner we will be able to make an impact with it.
  • The total global market for AI in 2016 was $4.06B. It is now projected to be around $169B in 2025. That’s some serious growth.
  • The amount of power you have to create solutions with this technology is immense. Artificial intelligence has been used in cancer research, gene therapy, brain-computer interface research, and everything in between.
  • If you aren’t interested in solutions, there are other motivations. The average machine learning developer makes around $114,121 annually. That puts the job in the top 10 highest paid jobs by average salary overall.
  • Most importantly: With AI at your fingertips, you will have access to so many opportunities to make a difference or achieve your personal goals that you would not have had before.

What's the problem?

  • The majority of the content available online is very difficult to understand. People explain concepts using a lot of jargon and don’t take the time to fully detail how concepts work at a basic level. This makes it extremely difficult to sift through resources to actually learn things yourself (without a Ph.D. level of understanding going into it)
  • None of the courses are fully comprehensive. This means that in order to get a broad overview of any topic, you need to jump around to a number of different courses just to learn about one topic in its entirety. This can get inefficient and time-consuming (and often causes us to lose our motivation)
  • A lot of the best content costs money. This is a big barrier to entry. If you are someone who is trying to learn about AI out of interest, or you are someone who doesn’t want to spend on this, the several hundred dollar courses might be a bit of an obstacle. There are some good free courses (which I will talk about later) but they are often hard to find.
  • No one tells you how you should be going about learning AI. This might be the most significant one. Learning to code and learning about artificial intelligence is not easy at all. It's one thing to know how to copy down code and pretend that you understand it. But to actually have a fundamental understanding and be able to create something from scratch yourself is far more difficult, and it takes time to get to that point. Also, no one is telling you how to get there.

I’m Adam Majmudar, a 17-year-old self-taught machine learning developer.

  • Connect with me on LinkedIn: My LinkedIn
  • Connect with me on Twitter: My Twitter
  • Follow me and see my other articles on Medium: My Medium
  • Send me an email:

What is this resource and how will it help

  1. Designed for Beginners: I have written everything with my audience in mind. I have made sure to explain everything to a level that anyone would understand, regardless of your experience coming into this.
  2. Understandable Content: I have attempted to make everything as understandable as possible, avoiding jargon when I can or fully explaining complex terms before I use them. My goal is to make all of this content accessible and easy to learn.
  3. Fundamentals Focused Content: I provide a detailed guide to understand AI, how it works, and what it can be used for.
  4. Creation Focused Content: I provide a detailed guide to learn how to create AI yourself (going into the details of learning to code + learning machine learning).
  5. My Experience & Tips: I provide tips and observations from my experience with artificial intelligence.
  6. High-Quality Resources & Guidance: I provide countless high-quality learning resources that are available online that will be the best for you to use to learn efficiently. When necessary, I also give my guidance on how to best use these resources to learn effectively. I’m also not trying to overwhelm you with external resources and am only providing content that I believe to be truly valuable.
  7. All Together & Comprehensive: One of the main advantages of this resource is that all of the information is condensed together in a convenient location. You don’t have to go searching across the internet to a million different websites to find the content that you need. Everything that you will need is either here, or linked here for you to refer to.
  8. Free (Mostly): Almost everything in this resource is free. Of course, this article itself is completely free. Almost every resource I will recommend is free as well. There are a few (very few) that cost money, but they are not necessary and are simply convenient for those that are willing to pay a small amount (I will also give tips on how to get these for the cheapest prices if you did want to purchase them).

Section 1: Gaining a Fundamental Understanding of Artificial Intelligence and Machine Learning

Fundamentals First

What is Artificial Intelligence?

This diagram shows the complexity of all of the topics that fall under the term of AI or are partially related (don’t worry if you don’t know what some of these things are)
  1. Narrow Intelligence: This is the type of AI that almost every (if not every) AI algorithm that exists today falls into. Narrow intelligence refers to AI that is designed to perform one specific task (no matter how complex that task is. AIs whose task is to talk to humans or to retrieve information from the internet or to play a board game are all examples of narrow intelligence.
  2. General Intelligence: This type of AI currently does not exist yet (and it might never exist). It refers to AIs that are able to perform a wide variety of tasks. In fact, they can perform any tasks that humans are capable of. This obviously requires a much more diverse form of intelligence and is much harder to achieve. It might even require the AI to be sentient (conscious and aware of themselves). However, for now, this is nothing but science fiction and a hope for the future.
  3. Superintelligence: This is a term coined by Nick Bostrom, the author of a famous book on AI called Superintelligence. He defines superintelligence as “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest.” Think something like SkyNet from the terminator. People frequently like to talk about this subject along with ethical questions about AI, but just like general intelligence, as of right now it is nothing but science fiction.

How Can Artificial Intelligence Be Used?

An Overview of Artificial Intelligence and Machine Learning

  • Algorithm: Set of rules and statistical techniques used to learn patterns from data
  • Model: A model is trained by using a machine learning algorithm
  • Predictor Variable: Feature of the data that can be used to predict the output
  • Response Variable (I will refer to this as Output): Feature or output variable that needs to be predicted using the predictor variables. In simple terms, this is what you are trying to predict.
  • Training Data: Machine learning model is built using training data
  • Testing Data: Machine learning model tested using the testing data
  1. Define Your Objective: What are you trying to predict? What are the target features? What is the input data? What kind of problem are you facing?
  2. Data Gathering: What kind of data is needed to solve the problem? Is this data available? How can you get this data?
  3. Data Preparation: Data cleaning involves getting rid of inconsistencies in the data set in order to make it suitable for computation. You are transforming data into a desirable format
  4. Exploratory Data Analysis: Understanding the patterns and trends in the data. All useful insights about the data and correlations are understood in this stage. This stage is very important in the machine learning process.
  5. Building a Machine Learning Model: Splitting your data into training and testing data to train your algorithm. You can choose from one of the many machine learning algorithms based on what’s best for the problem.
  6. Model Evaluation & Optimization: Testing the accuracy of your model and then tuning your model and optimizing it so it can predict what you want it to accurately. After you test your model, you need to find the accuracy of your algorithm
  7. Predictions: This is the step where you use your algorithm which has already been trained to come to valuable conclusions.
  • Supervised Learning: Technique in which we teach or train the machines using data that is well labeled.
  • Unsupervised Learning: Technique where the machine is trained using information that is unlabeled and we allow the algorithm to act on that information without guidance.
  • Reinforcement Learning: The part of machine learning where an agent is put in an environment and learns to behave in this environment by performing certain actions and observing the rewards which it gets from those actions.
  • Regression: Used for making predictions where the output is a continuous value (like a number). For example, predicting the price of something could be a regression problem.
  • Classification: Used for making predictions where the output is a categorical value. This simply means that you are predicting what category something belongs to. For example, determining if a picture is a dog or a cat would be a classification problem.
  • Clustering: Used for grouping similar things or data points into clusters. This would be like getting a bunch of pictures of cats and dogs and grouping the cats together and the dogs together (the algorithm doesn’t need to recognize which pictures are cats and which are dogs, it just needs to recognize which pictures are the same type of animal as each other. It might seem similar to classification but there is a subtle difference there.
Here is a helpful chart from the video showing different characteristics about the different types of machine learning and the types of problems that they can be used to solve.
  • Linear Regression (Regression)
  • Logistic Regression (Classification — weird right… why is it called regression but is actually classification??)
  • Decision Tree (Classification)
  • Random Forest (Classification)
  • Naive Bayes (Classification)
  • K-Nearest Neighbor (Classification)
  • Support Vector Machines (Classification & Regression)
  • K-Means Clustering (Clustering)

What are Neural Networks (Deep Learning)?

Here is a helpful chart showing the relationship between AI, ML, and Deep Learning

Section 2: Learning How to Create Artificial Intelligence Yourself

Introduction to Programming

Goals with Learning to Code

  1. Understanding and being able to use syntax, which is basically just the vocabulary of a certain programming language that you use to code.
  2. Being able to apply logic and critical thinking to solve a problem.

The Programmer’s Mindset: Figure Stuff Out (IMPORTANT!)

  1. Search it on Google.
  2. Search through Stack Overflow, and if you can’t find anything on your question, then you can ask a question yourself.

Getting Started with Python: A Step By Step Approach

  • Codecademy Python 3 Course (Pro) — If you are willing to pay for the convenience of learning Python 3, then by all means do. However, there is a free alternative (which is slightly less convenient but still effective)
  • Codecademy Python 2 Course (Free) — Alternatively, you can take this course on Python 2. The thing about Python 2 is that it is almost EXACTLY the same as Python 3, with only a few key differences. Therefore, if you take this course, you can read the following article:
  • Differences Between Python 2 and Python 3 — You can read this article to learn the differences between Python 2 and Python 3. That way, with a few changes to the syntax that you use, you will know how to program in Python 3.

Start Learning to Code Artificial Intelligence and Machine Learning

Section 3: A Deep Dive on Machine Learning and Deep Learning

Natural Language Processing (NLP)

We haven’t covered much about NLP yet, but I personally find it to be one of the most interesting and powerful forms of AI.

Reinforcement Learning

This simple cartoon captures all of the most important aspects of Reinforcement Learning
This video gives an amazing example of the potential of reinforcement learning.

Convolutional Neural Networks (CNNs)

Convolutional Neural Network: A Step By Step Guide - Towards Data ...
The basic process of a CNN (this will be explained in the resources I provide you). The idea here is that it is trying to recognize what is in the image by using an altered form of a neural network.

Recurrent Neural Networks (RNNs)

This is a simplified example of RNNs involving food that you will learn about in one of the videos below.


Autoencoders are a cool form of deep learning which can be used to manipulate images in different ways to get some useful results (in this picture, an autoencoder is used to get distracting dots on the images on the left to make the images clearer.)

Generative Adversarial Networks (GANs)

Here are some faces generated by GANs. Only the two photos in the left column were real photos that were actually taken. The remaining photos were completely generated by an AI algorithm that referenced those first two photos!

Where to Go From Here

1. Dive Deeper Into More Technical Topics

2. Focus on Content Creation

3. Learn About the Intersection Between AI & Another Field

A Quick Closing Note

Contact Me

  • Connect with me on LinkedIn: My LinkedIn
  • Connect with me on Twitter: My Twitter
  • Follow me and see my other articles on Medium: My Medium
  • Send me an email:



Love podcasts or audiobooks? Learn on the go with our new app.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store