Artificial Intelligence has emerged from the laboratory to the marketplace and its impact on society is growing rapidly. Thus, it is important for people to understand the fundamental of this fascinating topic. In this Article, I will explain you the differences between Artificial Intelligence, Machine Learning, and Deep Learning. I want to emphasize that these topics have been studied by experts for decades. The goal here is not to give you a deep understanding, but an intuitive understanding.
For a long time, we were trying to replicate our thought process by giving a lot of explicitly programmed instructions (rules) to computers. We thought that, as a result, we would come up with Artificial Intelligence. Needless to say that it turned out to be a terrible way. Instead, researchers began to write algorithms that take in data and learn from themselves. It was the beginning of Machine Learning.
Machine learning is a subfield of Artificial Intelligence where machines can learn from data without explicitly programmed instructions.
For instance, let’s suppose that we have to build an algorithm that can play tic tac toe. In the AI case, you would have to take it by its hand and give it many logical rules to learn how to play. As a result, you cannot build a machine that is good at tic-tac-toe without being good at it yourself. In the Machine Learning case, you may not know how to play tic-tac-toe yourself (yes, that is weird), yet you can still come up with a great software. Unlike the AI approach, you give it many examples of previous games, and let it learn the rules by itself.
You may have heard that Machine Learning is the current state-of-the-art. It is showing the most promise at providing tools that the industry and society can use.
Deep Learning is a subfield of Machine Learning. You can consider it as the cutting edge of the cutting edge. Like in Machine Learning, the data is fed through neural networks (algorithms that take inspiration from the human brain). Those artificial neural networks will extract a numerical value for every data (image, voice, text…) which pass through them and classify it.
Let’s suppose that we want to build an animal classifier that tells you if a given picture is a cat or a dog. In the Machine Learning case, we would have to define features such as if the animal has ears, and if yes, then if they are pointed. We would have to define all the facial features and let the system learn which features are the most important to classify a specific animal.
Deep Learning takes a step ahead and automatically finds out the features which are important for classification.
Deep Learning has only recently become useful, and there are many reasons to that. First off, deep learning requires a large amount of data which has been possible thanks to the recent digitalization of our society (everybody is now using a smartphone that produces data). Secondly, Deep Learning requires substantial computational power which is available today compared to previous years.
Let’s note that the human brain doesn’t need a lot of data to learn because it utilizes the knowledge from previous experience to accelerate learning from the future. Deep Learning requires many data because of the huge number of parameters needed to be tuned by a learning algorithm. The problem with Deep Learning is that it starts off with a poor initial state. And as you can imagine, requiring a large amount of data is a drawback, because it is hard to gather. Similarly, having substantial computational power is costly. Consequently, Deep Learning is expensive.
Is Deep Learning the way to go? Deep Learning pioneer Geoffrey Hinton says we need to start over. More about it here: https://www.axios.com/ai-pioneer-advocates-starting-over-2485537027.html
Thanks so much for reading, that’s the end of Article 100. Share it with your friends and don’t forget to follow me on social media. Subscribe to my newsletter down below to get your free eBook and learn the basics of Artificial Intelligence.
FOLLOW MY INSTAGRAM 😊- http://instagram.com/selimchehimi
FOLLOW MY TWITTER- https://twitter.com/SelimChehimi