Source: Deep Learning on Medium
Artificial Intelligence, Deep Learning, and Machine Learning
“How they’re different and why they are all essential to the “Internet of Things”
I had to go back and re-read this post. Not only are they fascinating topics to study, but they are also even fun to talk about.
I want to look at the explanations by Calum McClelland who is the Vice President of Operations and Projects at Leverage. He is also Director of Operations for IoT.
I found his article on Medium several years ago. I will try to bring it up to date.
Mr. McClelland starts his discussion by getting everyone on the same page. “We’re all familiar with the term “Artificial Intelligence.” It’s been a popular focus in movies such as “The Terminator,” “The Matrix”, and “Ex Machina.”
“I’ll begin by explaining what AI, ML, and DL mean and how they’re different.
He continues “then, I’ll share how AI and the Internet of Things are inextricably intertwined, with several technological advances converging at once to set the foundation for an AI and IoT explosion.”
He called that right. I read this article on February 2nd, 2017. Let’s try to expect what he predicted and what has been taking place in the last two years.
John McCarty, in 1956, first coined the phrase “Artificial Intelligence” or “AI”. AI can perform tasks characteristic of human intelligence. It includes planning, understanding language, recognizing objects and sounds, learning, and problem-solving.
“We can put AI into two categories, general and narrow. General AI would have all the characteristics of human intelligence, including the capacities mentioned above.
Narrow AI exhibits some facets of human intelligence, and can do that facet extremely well, but is lacking in other areas. A machine that’s great for recognizing images but nothing else would be an example of narrow AI.”
Machine learning is a way of achieving AI.
Arthur Samuel coined the phrase Machine Intelligence in 1959. He defined it as, “the ability to learn without being explicitly programmed.” You can get to AI without using Machine Learning but oh it would take building millions of lines of code with all the complex rules and decision-trees.
Instead of hard-coding software routines to execute a particular task, machine learning is a way of ‘training’ an algorithm so it can learn how to do it. “Training” involves dealing with massive amounts of data sets and allowing the algorithm to sort itself out by adjusting and improving.
Something like the Marine Corps teaches ‘Adapt and Execute.” At least it did when my three sons were in the corps. That was about oh 20 years ago. I guess that really dates me and believe me I feel every pulled muscle and strained ligament from each misstep during that time.
Machine learning has helped to improve computer ‘vision’ (the ability of a machine to recognize an object in an image or video).
Gather millions of pictures and have humans tag them. They would take a picture of a dog and of a cat. When the algorithm is instructed to build a model of a cat, it can tell the difference. When it has performed this feat accurately thousands of times, it has ‘learned’ what the difference between that cat and that dog.
Deep learning is one of the many approaches to machine learning. Other approaches include decision tree learning, inductive logic programming, and others.
The brain, its structure, and its function inspired deep learning. Also, the neurons are interconnected. Artificial Neural Networks (ANNs) are algorithms that mimic the biological structures of the brain.
“In ANNs, there are “neurons” which have discrete (individually separate and distinct” layers and connections to other neurons. Each layer picks out a specific feature to learn, such as curves or edges in image recognition.
It is this layering that gives deep learning its name, otherwise stated as “depth created by using multiple layers as opposed to a single layer.
Just before going to press with this article IBM announced they had made great strides in ‘speeding up’ these processes which will save incredible amounts of time to learn and get functioning.
Artificial Intelligence and the Internet of Things (IoT) are much like the relationship between the human brain and body.
“Our bodies collect sensory input such as sight, sound, and touch. Our brains take that data and make sense of it, turning light into recognizable objects and sounds into understandable speech. Our brains then decide, sending signals back out to the body to command movements like pickup up an object or speaking.”
This Internet of Things is like our body, sensors connected and raw data streaming in and forwarded to the “brain”. Artificial Intelligence makes sense of that data and deciding what actions to perform.
The connected devices of IoT are on the receiving side, carrying out the actions or communicating with other devices.
That is far enough today since we have other things to look at.