Generative AI For The Rest Of US

AI, Machine Learning, and Deep Learning: How Do They Relate?

If you are not familiar with the terms AI, Machine Learning, and Deep Learning, they might sound confusing, and for some, they might even sound like the same concept. They're not. There's a connection between them and, more specifically, a hierarchy. Here's a simple way to understand them:

First of all, AI is the broadest term. It refers to a field of computer science that focuses on the ability of a machine to perform tasks that typically require human intelligence: understanding natural language, recognizing objects in images, playing chess, driving a car... Now, to enable machines to perform these tasks, we need a tool, an engineering discipline, and there are many of them. Machine Learning is the most popular one today.

Machine Learning is not the only "tool" we have at hand, there are other tools, like rule-based systems where you have to write the rules that the machine will follow to perform a task. In contrast, in Machine Learning, the algorithm learns the rules from data. You give the algorithm a lot of examples of the task you want it to perform, and it creates its rules from these examples.

Machine learning can be categorized into different types, such as supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Each type differs in how it interacts with the data provided to it.

• Supervised learning involves algorithms learning from a labeled dataset. In this case, the correct output is provided, and the model is trained to predict the output from the input data. Supervised learning is similar to teaching a child how to differentiate between different types of fruits by showing them examples. Imagine you have a basket filled with various fruits: apples, bananas, and oranges. You take each fruit, show it to the child, and say its name aloud. "This is an apple," you explain as you hold up an apple, "and this is a banana," you continue, doing the same with a banana and an orange.

• Unsupervised learning involves algorithms learning from an unlabeled dataset. Here, the model is trained to find patterns in the data without being told what the patterns are. If we build on the previous example, unsupervised learning would be like showing the child the basket of fruits without telling them what each fruit is. The child would have to group the fruits based on similarities, such as color, shape, or size, without knowing the names of the fruits.

• Semi-supervised learning