Member-only story
What is AI? 4 Key Concepts for Beginners and Mathematicians (Part 1)
AI, machine learning, deep learning, neural networks: we see these words everywhere. Yet, differentiating between them is still confusing for many nonspecialists.
In this series of articles, I will attempt to explain each one simply. This would be accessible to non-technically minded readers. I will then provide a mathematical framework for the more technically inclined audience.
The Basics
In simplest terms, we have the following hierarchy:
- AI: simulation of human intelligence by machines and computer systems. The goal of AI is to enable machines to make their own decisions and predictions without needing explicit programming.
- Machine learning: subfield of AI that uses data and algorithms to achieve artificial intelligence in machines.
- Neural networks: sub-field of machine learning, where computer systems are modelled on the human brain and nervous system to aid machine learning algorithms.
- Deep learning: a sub-field of neural networks. To achieve deep learning, multilayered neural networks are used to identify and solve more complex classification problems.
Pictorially, we see the following diagram:
For the rest of this article, we focus on machine learning.
Supervised vs Unsupervised Learning
There are two main streams of machine learning: supervised and unsupervised. The main difference between them is the data we feed the machine to learn. Supervised learning uses so-called labelled data, whereas unsupervised uses unlabelled data.
Without getting too technical, labelled data is what you may intuitively think it is: a dataset where everything is clearly categorised, or labelled. This aids the machine’s learning process.
