Clearing up the Confusion – AI vs Machine Learning vs Data Science
The buzz words these days are artificial intelligence (AI), machine learning (ML) and data science. Everyone is realising the benefits that these things offer and are eager to jump on the band-wagon. Unfortunately, there are a lot of misconceptions regarding these topics, and in this article, I will try to clear up the confusion.
Let’s start off with Data Science. This is the study of extracting insights from data using the scientific method. This is nothing new; the least-square regression method has been around since the 1800’s. The difference is, now we have the technology and vast amounts of data to use these methods effectively. Think of statistics and advanced analytics.
Artificial intelligence is often used synonymously with machine learning. However, it is much broader than that. AI can be described as the development of computer systems to perform tasks that normally require human intelligence; visual perception, speech recognition, translation between languages and decision making. Think of iRobot, or Terminator.
Machine learning, on the other hand, is a means to get machines to perform certain tasks, without using explicit instructions, but rather, being trained from large amounts of labelled data.
Think of it like this: imagine you want to ‘teach’ a machine learning algorithm to learn the difference between pictures of dogs and buses. If you show the machine thousands of images of buses, labelled “bus“, and thousands of images of dogs, labelled “dog“, the algorithm will find similarities in the pictures of buses and associate them with the label “bus“, and it will find similarities in the pictures of dogs and associate them with the label “dog“. Then, when you show the machine a picture of a new dog it hasn’t seen before, it will look at the picture and compare it to the thousands of images it has seen previously. It will then see the similarities to the label “dog“, and it will think “Hey, this looks familiar. I think this is a dog”. How it actually goes about this is the subject of another article, but it is fascinating.
So, why would someone want to teach a machine to distinguish buses and dogs? Well, for one, this is the sort of thing you would want to train your self-driving car to avoid. But the number of questions answerable by machine learning is limited only by your imagination. Will this employee stay with the company or resign? Is this pattern of purchasing significantly different for this customer? What aspect in the market will dictate a new fashion trend? Can wearable blood-sugar level monitors predict a diabetic sugar crash before it happens? Machine learning is in many things you use daily, without you ever realising it. The spam filter on your e-mails, the recommendations on Netflix and YouTube or the bank calling you about a suspicious transaction when you transfer a large amount of money.
The world and workplace are full of opportunities, and data science, machine learning and artificial intelligence are ready to help you get the most out of them. All you need is some data and creativity.