Artificial Intelligence vs Machine Learning vs Deep Learning
In today’s technological society, we are benefitting from artificial intelligence every day such as Google Maps, Uber, music recommendations, and other applications that are powered by artificial intelligence. Over time, there has been a confusion in terminology between artificial intelligence, machine learning, and deep learning when speaking about computers thinking independently. In this post, I will define the three different terminologies and their differences.
Artificial Intelligence
Artificial Intelligence is the science behind building intelligent programs or machines that can solve problems independently. The term was first coined in 1956 at a computer science conference in Dartmouth. Scientists expected that to understand how the human mind works and to digitalise it, it should not take that long. It’s now 2020 and we are still developing the technology.
Artificial Intelligence can be divided into three categories:
Artificial Narrow Intelligence
Focuses on a particular task and can not pass as human due to the machine being created for a specific purpose. For example in 1996, a computer called Deep Blue became the first computer to beat a human at chess. Deep Blue was able to generate and evaluate 200 million chess positions per second.
Artificial General Intelligence
The computer becomes more human-like, being able to make their own decisions and learn without any human input. The computer can also show emotions and can maintain a conversation.
Artificial Super Intelligence
This is the type of machine that is the general vision that everyone has when speaking about artificial intelligence. This is where the computer is way ahead of humans; smart, wise, creative, and has excellent social skills. The goal of the machine is to make human’s lives better or destroy them all. This vision is often depicted in films such as The Terminator franchise where a computer causes a mass attack on human civilisation.
Machine Learning
Machine learning is the subcategory of artificial intelligence that focuses on teaching computers how to learn without the need to be programmed for a specific task. To enable machines to learn you need 3 components:
Dataset
Machines are trained on special collections of data samples which could include numbers, images, texts or any other kind of data.
Features
Machines need to know the important pieces of data that work towards the solution of the problem. Without it, machines will not know what problem they are solving.
Algorithm
An algorithm is a piece of mathematical equations/code to help solve the problem. You can solve the same solution using different algorithms. Depending on the algorithms you may get to the solution quickly and the accuracy of the result is also dependent on the algorithm.
Deep Learning
Deep learning is the type of learning algorithms that are based on the structure of a human brain. Deep learning uses complex multi-layered neural networks where information is transferred from one layer to another, similar to a human brain until the output is generated. To train neural network, scientists need a huge amount of training data as there are a large number of parameters that need to be considered for the best solution output. Deep learning has several applications such as image recognition, drug discovery, speech recognition, and many more. Deep learning is the trend when developing artificial intelligence.
Summary
Artificial Intelligence will be an ongoing development where with each technological advancement, artificial intelligence improves and becomes smarter. There are many variables to consider when developing a machine that can think for itself. Hopefully this post has provided an insight into the terminology and the differences between them.
Check out What is AI? blog post if you want a summary in artificial intelligence.
Video: Ironman using jarvis
[youtube=://www.youtube.com/watch?v=ZwOxM0-byvc&w=480&h=270]