What’s what: Artificial intelligence, machine learning, and deep learning

Artificial intelligence (AI) and machine learning (ML) are terms being used in various industries to explain their latest foray into next-generation technology. Add deep learning (DL) to the mix and things start to get really confusing.

While AI and ML can be used interchangeably in many contexts, there are some serious differences between them. For businesses looking to embark on exploring new AI solutions, distinguishing one from the other is critical to identifying which your business needs and what can help it the most.

Artificial Intelligence

The ability for a machine to mimic human intelligence to perform human-like functions.

Everything from Apple’s Siri voice assistant to self-driving cars falls under AI.

AI is an umbrella term for any machine that exhibits any form of human-like intelligence.

However, there are many ways to achieve AI – you can have a thousand engineers code millions of lines of code, setting the parameters for every possible function, or you can have the computer learn on its own.

So far, a lot of the AI in the market are coded intelligence, where every outcome is written out by humans.

If this was math: You are writing basic sums. Like 2 + 3 = y, and you need to find Y.

This is simple, straightforward and you’re dealing with only numbers. This type of equation is the building block for the many types of calculations available in math.  You know all the numbers by heart and how to use them.

Machine Learning

One of the many ways for a machine to acquire human-like intelligence.

This is where “the computer learns on its own” comes in. Instead of coding every possibility, you “train” an algorithm, by feeding the computer with a bunch of data so it learns and adjusts its functions.

It includes things like visual recognition and recognizing spam emails. The computer sees the same patterns over and over, and learns over time, that a particular picture is a cat, or a mail is a spam, so kick it to the junk.

If this was math: You are finding the length of the sides of a triangle. Figure out the right function you need to find that length. There are pictures involve now but you are still doing basic sums. You need to take into account how the angles affect each other and create a sum that helps you find the length. You are probably the expert in triangles now.

Deep Learning

One of the many branches of machine learning.

In short, this is engineers trying to build a brain. Inspired by how the brain works, “Artificial Neural Networks” (the fancy name for deep learning) is designed to mimic connected neurons. Engineers run many layers of algorithms, for the machine to perform several functions, collectively.

This is how they train self-driving cars and why it’s taking so long. The cars have to recognize obstacles, drive, switch lanes, turn corners, adhere to traffic rules.

Although the image that the car “sees” is just a road, with trees or other objects on it, there are many functions it needs to run. It will need to understand what all the different outcomes mean, then putting it all together to decide whether it should stop or drive or turn.

If this was math: You are trying to build a house. You need many different calculations to figure out the dimensions.

On top of that, you need to understand the properties (such as the mass) of cement or steel or clay bricks, and how that will change your designs on a paper.

This requires more thought and involves more than numbers and pictures, but the end calculation still boils down to basic algebra. You are now a building architect.

In a nutshell, AI is the umbrella term, ML is one of the ways to achieve AI, and DL is an implementation of ML.

You can use AI right now if you know how to code a certain function you need for your business. ML and DL require more specialty and it’s best to partner with ML companies, who can advise you how to design and build that “house”.

AI, ML and DL are all tools to help businesses. Just as Word allows you to type documents, it is up to you how you craft the story using the tools provided on Word. How these “equations” can come together to suit your business is down to the design and implementation of each system.