How do I learn the basics of AI?
In my previous article in this series on Getting Started with AI, we discussed university degrees in AI.
In this article, we’ll learn about the basic AI skills that you need to know for a career in AI.
Based on the career path that you’ve chosen, there is an ideal curriculum for learning basic AI skills.
All of these learning paths require basic AI literacy.
However, some require more advanced AI knowledge and skills.
In this article, we’ll discuss how to learn the basics of AI for each learning path.
For every learning path, basic AI literacy is an essential skill. There are a lot of concepts that you will need to know. For example, the difference between narrow AI (ANI) and general AI (AGI). How machine learning, deep learning, and reinforcement learning work. What jobs are at high risk vs. low risk of automation from AI? How to invest in an AI-driven economy. How to use AI responsibly and ethically.
To learn basic AI literacy, I recommend starting with my free online course Preparing Your Career for AI. It will teach you the five things that everyone should be doing today to prepare for the coming wave of AI automation. Next, I recommend my online course Artificial Intelligence: The Big Picture. It provides a high-level overview of the entire AI space. To dig deeper, I recommend Andrew Ng’s course AI for Everyone.
For most learning paths, you will need to learn a programing language like Python. If you want to call pre-trained AI services via REST APIs, then you can use any language you want. However, if you want to train or customize models, I recommend learning Python plus R and SQL. Then learn the various Python AI/ML packages like Numpy, Pandas, Scikit Learn, Pytorch, Keras, and TensorFlow.
To learn Python, I have a few recommendations. For beginners, I recommend DataCamp or Codecademy. They provide interactive tutorials to guide you through the basics of programming. If you already know how to program, you might want to try Pluralsight courses — which are designed for IT professionals. For expert programmers, I recommend Real Python. It allows you to learn at a much faster pace.
If you want to train or customize models, you will need to know Machine Learning (ML). A wide variety of tools and techniques are used to train, tune, evaluate, and deploy ML models. You will need to know about supervised, unsupervised, reinforcement, and deep learning. You will also need to know how to train and use decision tree classifiers, k-nearest neighbors, logistic regression, SVMs, and neural networks.
To learn ML, I have a few recommendations. If you are a complete beginner, I recommend starting with DataCamp, Kaggle, StatQuest, and 3Blue1Brown. If you already know a bit about ML, I recommend courses on Pluralsight, Coursera, and edX. Next, check out my course on Deep Learning: The Big Picture. Then, start playing around with models using ML model repositories like HuggingFace.
If you want to train or customize new models, you will also want to know some basic Data Science (DS). DS is the foundation of modern data-driven AI. It’s a set of practices for turning data into actionable insight. This includes creating, collecting, cleaning, transforming, analyzing, and visualizing data. It also includes working with Big Data, training ML models, and analyzing their performance.
To learn DS, I have a few recommendations. If you’re a complete beginner, I recommend my free online course Intro to Data for Data Science. Next, take my online course Data Science: The Big Picture. Then, watch videos by StatQuest and ritvikmath or take courses on DataCamp, Pluralsight, or Coursera — listed in order of difficulty. To dig deeper, I recommend a Data Science Specialization from JHU.
ML is just one part of Artificial Intelligence (AI). However, when most people say AI, they typically mean ML. Beyond ML, there are a variety of “algorithms with tricks” that are used to solve complex problems. While you can go far in AI using only ML, other techniques are often necessary. These techniques include heuristic search, symbolic logic, expert systems, constrained optimization, genetic programming, swarm intelligence, etc.
To learn (non-ML) AI, I recommend you start with my online course Artificial Intelligence: The Big Picture. It provides a high-level overview of modern data-driven AI. After this, I recommend that you read Artificial Intelligence: A Modern Approach. It’s the most popular textbook on AI. It’s a bit technical and academic, but other than some math and statistics, it’s still accessible to most developers and IT professionals.
Knowing about AI is one thing. However, knowing how to solve problems with AI is an entirely different skill set. So, you will need to know a wide variety of tools and techniques to solve various problems that can only be solved with AI. You need to be able to look at a problem and immediately know how you might solve it with AI or at least know where to start looking for a potential solution.
To learn how to solve problems with AI, I recommend checking out my online course The AI Developer’s Toolkit. It covers the entire landscape of problems and solutions in the AI/ML space. It will teach you how to apply AI to analyze and synthesize tables, text, audio, images, and video. In addition, it will show you how to compose AI applications and build AI systems.
Sometimes it’s possible to solve a problem with AI; however, there isn’t a business case for doing so. Business aspects (beyond the technical aspects) impact whether AI is the right solution for a specific problem. For example, cost-benefit analysis, return on investment (ROI), competitive advantage, economies of scale, and network effects. It’s important to learn how these concepts apply to AI.
To learn AI business strategy, I recommend you watch my presentation on Developing Your AI Strategy. It teaches the key concepts of AI business strategy with case studies of AI transformation successes and failures. If you want to dig deeper, I recommend Andrew Ng’s online course on AI for Everyone or MIT’s more expensive course on Artificial Intelligence: Implications for Business Strategy.
AI is a powerful technology, and it’s getting more powerful every day. However, most of us have not yet learned how to use our new AI “superpowers” responsibly and ethically. As a result, you need to understand the ethics of AI to avoid the serious legal, ethical, and social consequences of AI on your organization.
To learn AI Ethics, I recommend you check out my presentation on The Ethics of AI. Next, you should take the free online course The Ethics of AI from the University of Helsinki. Then, I recommend you look into Microsoft’s Responsible AI Principles and Google’s Responsible AI Practices. Finally, to go deeper, read The Age of Surveillance Capitalism, AI Ethics (MIT Press), and The Alignment Problem.
The technologies fueling the AI revolution are changing fast. It’s hard to keep up with all of the new tools, techniques, research, and models, coming every day. In fact, it seems like there is a big breakthrough in research and practical applications happening every few days.
To keep up with the latest trends in AI research, I recommend watching videos by Two-minute Papers, Matt Wolfe, and ColdFusion. They cover the latest AI news and research in short videos. I also like FutureTools, Arxive Sanity Preserver, and Papers with Code for the latest AI tools and research. I also follow AI research from Microsoft, Google, Open AI, DeepMind, and NVIDIA. Finally, I recommend HuggingFace as my go-to resource for the latest in open-source AI models.
Here’s a list of all the best free resources I’ve found to learn foundational AI/ML concepts. I’ve spent a lot of time searching for the best teachers on these topics, and these are the cream of the crop. https://matthewrenze.com/articles/best-free-resources-to-learn-ai/
To learn more, please read the next article in this series: Learning Practical AI Skills.