Artificial Intelligence – A Practical Primer

Introduction

Artificial Intelligence (AI), machine learning (ML) and deep learning (DL) seem to be taking the world by storm, allowing simple creation of things that were previously, digitally impossible. No sector appears to be safe.

But does the hype exceed reality?

Does this mean we will all be programming neural networks? Or will every business be replaced by one of the big tech companies or a even fresh start up?

In this blog post series, we will explain what Artificial Intelligence, Machine Learning and Deep Learning are and what they could mean for you and/or your organization.

Artificial Intelligence (AI) has been a rising star, as well as having encountered some ‘winters’ for not delivering on expected value. These days with Deep learning, AI is back on the rise, but also subject to being (over)hyped: the AI robots are taking over, it’s taking everybody’s job, or will solve all problems by itself…

…so, in this “practical primer” we will keep our heads in the clouds to explore all the possibilities AI can offer, what and how it can actually solve real world problems for you…

…but we are pragmatists, we deliver what works, and research practical aspects of AI.

So, at the beginning of this series, we make one promise:

We promise to keep both feet on the ground as to what AI can actually do, and how you can apply it practically.

We will take you from definitions to how you can apply AI in your organization, keeping it practical, but not shy away from the details where they are useful. A spoiler, it’s never rocket science…

We realize that this is a lofty goal, but what we are writing here is something we shared with over 100 companies in their search to implement AI.

This post will serve as an overview, while we will update the links as we go.

Table of contents

1. Definitions
   1. Artificial Intelligence
   2. Machine Learning
   3. Deep Learning
      1. Computer vision
      2. Natural Language
      3. Structured Data
      4. Some special topics?
2. Why is this working now?
   1. Lots of Data
   2. Massive Compute
   3. Improved Algorithms
   4. Accessible Frameworks
3. Myths about AI
   1. Do you need a Phd for AI?
   2. Do you need big data for AI?
   3. Do you need big compute for AI?
   4. Will AI replace domain experts?
4. Learn from experience
   1. Supervised Learning: From labelled examples
   2. Unsupervised Learning: From the examples themselves
   3. Reinforcement Learning: Interact with + get feedback from an environment
5. A more detailed look at supervised learning
6. Transfer Learning
7. So, that’s settled. Where do we begin?
8. When not to use AI?
9. Choose your metric
10. Choose your AI superpower
11. It’s about the data, stupid
12. Build vs Buy
    1. Buy: Get get things done (for solved problems)
    2. Build: You know you want to
13. Should I train on my laptop? A journey through platforms.
14. If an AI model cannot be used, does it exist?
15. There’s practical and then there is practical practical…
16. AI use case: from start to finish
17. AI Use cases