ARTIFICIAL INTELLIGENCE, DEEP LEARNING
What is Artifical Intelligence, Machine Learning and Deep Learning?
Have you ever been confused by the terms Artificial Intelligence, Machine Learning, and Deep Learning? Every time a new concept or technique is invented, a new term is introduced to the public. As a result, the difference between all terms can be very unclear and they can be used either interchangeably or incorrectly. In the case of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) the concepts are closely related which makes it even more challenging to highlight the differences.
The easiest way to think of their evolution is to imagine them as concentric circles where AI is the idea that came first, therefore, it’s the largest, overarching concept. AI stands for human intelligence exhibited by machines. Then Machine Learning came later as an approach to achieve AI. In recent years, Deep Learning — which can be considered a technique for implementing ML is experiencing its worldwide explosion. In this article, we will cover the basics of all these concepts to give you a better understanding.
What is AI?
Artificial Intelligence was founded in 1956 as an academic discipline. Back then, the goal was to get computers to perform human tasks: mimic the human decision-making process.
In the beginning, researchers developed and experimented with computer programs that could play checkers and solve logic problems. Looking at the output produced by the programs playing checkers, researchers could see some form of artificial intelligence behind the program’s moves, particularly when a human player was beaten. Based on this simple explanation, we can say that artificial intelligence refers to the output of a computer because the computer is doing something intelligent. It is the simulation of human intelligence processes by machines, especially computer systems.
What is Machine Learning (ML)?
In the early days as researchers were progressively building the foundation for computers with AI, they also encountered some problems. They found that some problems were much harder for the computer to solve. This was because the problems were not amenable to the early techniques used for AI. Researchers learned that to overcome these problems, computers had to not only mimic human behavior but to mimic how humans learn as well.
For a computer program to learn, it needs to be taught. The idea behind ML is therefore centered around the ability of a computer program to process lots of data and be able to learn the rules and exceptions pertaining to a particular task. For example, feed a computer program (or an algorithm) lots of data on financial transactions, tell it which ones are fraudulent, and let it work out what indicates fraud so it can predict fraud in the future.
Further exploration into ML lead to researchers going all the way, beyond mimicking human learning. The idea was that if machine learning is about mimicking how humans learn, why not mimic how the human brain works. This lead to the development of artificial neurons. Neural networks simulated by a computer program started being used to solve certain problems.
VIDEO: New AI Imaging Technique using deep learning framework reconstructs photos with realistic results
What is Deep Learning?
Deep learning is a subset of machine learning and a specific approach used for the building and training of neural networks. A deep learning model is designed to continually analyze data with a logic structure similar to how a human would draw conclusions. To achieve this, deep learning uses a layered structure of algorithms called an artificial neural network (ANN).
Deep learning differs from traditional machine learning techniques in that the input data is passed through a series of nonlinearities or nonlinear transformations before it becomes output. To further explain, a machine learning algorithm often needs an engineer to step in and make adjustments if it returns with an inaccurate prediction. With a deep learning model, however, the algorithms can determine on their own if a prediction is accurate or not.
This is just a simple explanation of the terms. If you want to learn more, click here for a more detailed and elaborate explanation.
Click here to read the full article.