Welcome to Knowledge Bytes, a series of short introduction articles to a topic surrounding data science, business and software development. Our goal is to make complex matters such as machine learning algorithms understandable in less than 5 min.
AI and ML
Artificial Intelligence (AI) and Machine Learning (ML) are certainly one of the hottest buzzwords in the tech industry today. These technologies will most likely shape our future more than any other technology this century. More so, while ML has been around for some decades, the technological advancements in computing power are changing the game and enable algorithms that we could not have dreamt of before.
The buzzwords explained
Artificial Intelligence and Machine Learning are often used interchangeably, but they are not the same thing. So let us clear this confusion.
While there are many definitions of AI, it basically comes down to a system or machine that is able to perceive the world around them, make plans and perform tasks therefore mimicking human intelligence. AI can be divided into general AI and narrow AI. Currently, humans are only able to produce narrow AI systems, meaning that this system behaves intelligently at one specific area, like ordering pizza. General AI is closer to human intelligence, a system that is knowledgeable in many fields and connect information from different fields to make better decisions and create new connections.
There are many fields that fall under the umbrella term of AI, like machine learning, natural language recognition or computer vision.
Machine Learning is therefore a subfield of artificial intelligence and was defined by Arthus Samuel (1959) as follows: “Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed.”  This means that instead of defining clear rules by which an application makes decisions, it identifies patterns by observing a set of data and therefore it figures out rules by itself.
An example of a simple Machine Learning application is a spam filter. Such a system can flag spam emails by learning from a set of spam emails (flagged by humans) and regular (non-spam) emails. This set of data is called a training set and by constantly giving the algorithm more data, it keeps on getting better. This is why, nowadays, you rarely find spam emails in your inbox.
Machine Learning in itself can be further divided into supervised-, unsupervised-, deep- and reinforcement learning. In the next 4 weeks, we will go into more detail and introduce some basic concepts about the different types of ML. Make sure to follow us on LinkedIn so you don’t miss an article!