Artificial Intelligence, Machine Learning and Deep Learning: A Primer.
by Dr Dorel Iosif
The field of artificial intelligence skyrocketed in popularity, and the sector is expecting accelerated growth rates in the coming years.
But many people have trouble articulating what the term means, and they often use AI, machine learning or deep learning interchangeably.
Read on for your introduction to these fields and the differences between them!
What Is Artificial Intelligence?
Darmouth College Workshop, July-August 1955. The term “artificial intelligence" is coined by John McCarthy (Dartmouth), Marvin Minsky (Harvard) and Nathaniel Rochester (IBM), initially in a proposal for a 2-month, 10-man study of artificial intelligence.
The proposal written in September 1955, stated:
“We propose that a 2-month, 10-man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.”
And the rest is, as they say, history.
Artificial Intelligence, or AI, has been a favourite term in academic and popular circles for decades. Unfortunately, the industry has struggled to maintain a strict definition of AI.
Basically, AI is a computer science that uses computers to solve problems analytically. It's a simple definition, if only because AI at its heart is a simple concept. It's the idea of creating processes that can solve problems without needing humans to perform actions every step of the way.
Many AI functions in artificial intelligence technology are simple series of conditional programs. These include things such as if-then statements. AI can help a company's innovation strategy or find complex statistical models.
First, a predetermined set of parameters is programmed in. The computer then follows the path of these statements based on the data it has until it arrives at its conclusion. In many ways, it makes the same decisions and conclusions a human would, simply at much faster speeds.
There are an astonishing range of applications of artificial intelligence.
But what about machine learning? How does it fit in with the overall picture of artificial intelligence?
What Is Machine Learning?
Although people often use the term machine learning interchangeably with artificial intelligence, there are differences between them. The most important thing to realize is that machine learning is a subset of artificial intelligence. By most definitions, all machine learning is AI, while only some AI incorporates machine learning.
The essential task of machine learning is to analyze data and find patterns or classifications.
A machine learning program is only as good as the data given, which is why it's vital to have good data, variables, and labels when using machine learning programs. The effectiveness of machine learning algorithms (mostly based on statistical methods) is directly proportional with the size of data, be that images, words, phrases, numbers. In other words, the larger the data sample is, the most effective the algorithms are.
Computers can use various machine learning algorithms to analyze data. They then create a model based on any patterns they see in the data, such as correlations between variables or changes under certain parameters. Data scientists, analysts, and statisticians can then use these models to create predictions and test the data and models further.
These machine learning algorithms don't necessarily have to be complex. A scientist could do something as simple as classifying people as "healthy" or "unhealthy". The data would then feed through a machine learning algorithm, which would create a model to predict a person's health based on other factors.
Machine learning models are used in a wide variety of fields and applications, from statistics and data science to email spam filters, and recommendation systems.
Simplistically put, machine learning needs: datasets, features and algorithms.
Another term that is often used in conjunction with machine learning is "deep learning". In the same way that machine learning is a subset of AI, deep learning is a subset of machine learning.
Specifically, deep learning utilizes neural networks with three or more layers. A neural network is a complex computer algorithm that contains many layers of nodes. These networks are made to mimic the human brain. The level of abstraction in the neural network increases gradually by non-linear transformations of input data.
Deep learning networks use layers of neural network nodes to analyze, classify, and predict data. Over time, these networks learn and improve, giving better results.
Deep learning (DL) requires a large amount of computing power and data. It is often more complex than is necessary for some situations. However, it is a powerful tool when used in the right applications.
The most advanced field in deep learning is called reinforcement learning or neuro-dynamic programming. They are best used in the long term vs short term reward tradeoffs. The goal of such models is to maximize total reward based on the exploration and exploitation of current knowledge. These DL models are based on a Markov decision process where decisions are made based on outcomes that are partly random and partly under the control of the decision maker.
Artificial Intelligence vs Machine Learning
Machine learning and artificial intelligence aren't as interchangeable as the way some people use them. However, neither are they as separated as some people like to define.
The basic difference is that machine learning is a subsection of artificial intelligence. It's the part of AI that uses training data to help improve and adjust its output models. It does this without the need for humans to monitor every single step of the process.
AI can be rather broad and vague in many ways, but is often focused on the process of coming to conclusions and making decisions. Machine learning is also focused on making decisions. However, it is more focused on making mathematical or statistical decisions to help analyze data.
How Are They Used Together?
Machine learning and AI have a lot of practical applications. When a data set needs analysed, scientists often turn to the team of machine learning nestled inside the field of AI. Whatever application they might have, the two fields can work together to give them the best result.
Someone might use a simple AI program to help make decisions on the best path to take with a business model. Or, they might run something more complex to help plan their business investment strategy and growth over the next quarter. Machine learning can help analyze the numbers and develop a solution that maximizes the chosen parameters.
These models and methods will work, but only within the parameters that are set. That's why it's so vital for computer scientists and analysts to use accurate data and parameters.
Learn More About Computer Technology
Now that you have a better understanding of the similarities and differences of artificial intelligence vs machine learning, you'll be more prepared when the subject comes up.
But if you'd like to learn more about these exciting fields of computer science, or you're interested in hiring experts to help you with your own projects, contact us!
We're eager to help you with all the technology aid you need to succeed and grow in these exciting times.