Both artificial intelligence and machine learning are hot buzzwords right now.
It's not surprising at all, since that’s exactly what modern technologies, such as: virtual agents, decision management, and even content creation, are starting to rely on.
While these two terms sometimes are used interchangeably, they shouldn’t be seen as one thing. In fact, some distinct differences that make them separate, and this is exactly what we’ll discuss today.
Movies such as “Ex machina,” “Her” and the “Westworld“ series on HBO are in on the hype, and we’re thrilled by the inevitable expansion of the technology into our reality.
Who can predict how far it will go?
Will living and breathing creatures eventually become redundant in keeping the system up and running?
One day we will see.
But before this happens, let’s try to explain these two complex terms in simple words and determine the differences between them.
In the past, even early European computers were considered by engineers as “logical machines” and “mechanical brains” because they were able to reproduce arithmetic and memory.
Along with progress in technology, our understanding of how the human mind works has evolved, and the concept of AI has changed.
The artificial intelligence we know today was invented by John McCarthy in 1956.
Simply put, AI is wrapped around mimicking human decision making processes and performing complex tasks in a more human-like way than ever.
It involves machines that have the ability to carry out tasks characteristic of human intelligence. It’s a much broader concept than machine learning.
These tasks include:
recognising voices and images,
and any tasks that would be considered “smart”.
There are two types of artificial intelligence. One is general, and the other is weak, or in other words, applied.
Weak AI is focused on one narrow (predefined) task. It can't generalise to unseen tasks and, of course, it doesn't have consciousness. As a result, it is characterized by excellence in that one capacity, but is lacking in other areas.
Generalized AI is representative of all the above mentioned capabilities, and also characteristic of human intelligence at the same time.
Weak AI is much more common and we can easily observe examples of its implementations around us. Weak AI systems include:
The Spotify discovery mode,
General AI is the intelligence of a machine that could successfully perform any intellectual task that a human being can. AI can make a computer smarter and more aware of its past iterations, so it can gain new capabilities and knowledge.
Systems or devices that are based on general AI and can perform any task don't exist in our reality yet, and it will a lot of time before any are created. However, this is where new advancements would be most exciting.
“Worrying about evil-killer AI today is like worrying about overpopulation on the planet Mars”, says Andrew Ng, VP and a former chief scientist of Baidu.
Fortunately, the type of AI we see in Westworld, or Ex Machina, is still a way off.
While Artificial Intelligence is a broad term, machine learning is a subset of AI algorithms.
Machine learning is a way of “learning” which enables an algorithm to evolve. In this case “learning” means feeding the algorithm with a massive amount of data so that it can adjust itself and continually improve.
The term “machine learning” was first coined by Artur Samuel, an American pioneer in computer gaming and artificial intelligence, back in 1959. Samuel defined it as “[The] Field of study that gives computers the ability to learn without being explicitly programmed”.
Machine learning lets computers recognise patterns in enormous datasets and act on them.
This is the current application of AI, and the execution of an idea of giving machines access to the data and letting them learn by themselves.
For instance, machine learning is exemplified by the ability of computers to recognize images.
Remember that Silicon Valley episode?
Let’s say you collect and input a massive number of different pictures into a system → and we’re talking about hundreds of thousands or even millions of pictures.
Some of them show hot-dogs.
Some of them show other foods.
We have tags for 50% images that say ‘hot-dog’ or ‘not-hot-dog’.
We don’t have tags for the other 50%.
A machine learning algorithm studies those images along with their tags, and finally it can build a model of a hot dog, which can later be used to classify pictures without any further description or tags.
This is how a machine learns to distinguish hot dogs from other foods.
Remember, artificial intelligence is a broad term that represents the general concept of machines being able to carry out smart tasks, and machine learning is a specific subset of algorithms for AI.
Still, these two terms have something in common. Both artificial intelligence and machine learning serve one goal.
They provide us with the unbelievably advanced technology that our grandparents never dreamed of, nor expected to happen.
Thanks to them, we can enjoy intelligent houses, cars, or even technology like Siri, that are becoming more and more helpful. And as long as we can benefit from it → it’s all great.
However, history proves that almost everything can be used in both good and bad ways. Again, technology can serve as another great example for one of those lessons.
How did you like this article? Let’s discuss the topic further. Speak your mind in the comments below.