Data Science, Machine Learning and Artificial Intelligence – what’s the difference?
Data Science, Machine Learning and Artificial Intelligence are terms that we often use when speaking to our clients and it's clear that we aren’t the only ones speaking about them – they seem to be buzzwords which everybody talks about today. Unfortunately, these words easily become muddled and get labelled as technical jargon causing confusion amongst non-technical professionals, and dare we say it, even technical professionals! We often find ourselves discussing and explaining the differences to business leaders who are beginning to harness the insights from their data sets and require the relevant professionals to do so. This bulletin aims to demystify the differences between Data Science, Machine Learning and Artificial Intelligence.
Data scientists use scientific methods to gain insights from data relevant to a business. Data Science lies at the intersection of computer science, mathematics, statistics and business, and is a very broad term. It has certainly progressed over the last decade given the increase in computing power that has led to the ubiquity of data, however, a significant portion of the excitement and developments of data science has been through finding applications of existing techniques on new data sets, rather than a new shift in the scientific paradigm. This is why there is a significant increase in demand for data scientists, as a wide-array of disciplines start applying mathematical techniques to their industries (see previous bulletin).
Machine Learning is a subset of data science and is the process by which a computer system is programmed to read and learn from a data set to provide insights and predictions for future outcomes. There are many types of machine learning algorithms and associated libraries which allow us to produce the desired output. The skill of a Data Scientist involves understanding the strengths and weaknesses of the algorithm and then working out the best way to modify the algorithm or if needs be, fine-tuning the relevant steps in the data processing cycle to produce the desired outcome.
As a simple example, imagine an online travel agent that wants to offer its customers a prediction of the weather for the dates they are looking to book a holiday. Having this information before booking can influence a customer’s decision to book or if given after booking it can be as part of a helpful guide of what to pack. This travel agent can provide this information by creating an algorithm to predict what the temperature will be for a given city in the future. With limited data (e.g. last 12 months of temperatures), the algorithm will perform far worse than if it had more substantial data (e.g. longer history e.g. 10 years, temperatures of neighbouring cities, weather patterns etc.). In general, the learning of the algorithm will only be as good as the quality and quantity of data.
Artificial Intelligence (AI)
Unfortunately, this is very much a term used to glamorise much of the work involved in data science. Whilst technical definitions will allude to computers demonstrating human like capabilities, we would argue that artificial intelligence could be, on the whole, simply viewed as nothing more than a narrative. In fact, one way to think of AI is that it not an actual thing. Instead, it is just how we describe the many actions produced as a result of the use of other technologies. These actions are comparable to human actions such as recognising objects, learning, understanding language and problem solving and therefore the term artificial intelligence has become synonymous with computers producing actions that exhibit elements of human intelligence. From our experience, when clients mention they need a new staff member to work on an AI project, extensive clarification is needed to truly understand the skill requirements for that project.
So how do they all link?
The differences and links between Data Science, Machine Learning and Artificial Intelligence can often be confusing but it doesn’t need to be! Let’s first cement the following two points:
Side Note: Artificial Intelligence, by definition, does not necessarily involve the use of data by computer systems.
Next is where the waters become murky; how does Machine Learning fit in with the above? Machine Learning can be considered as the link between the two. Consider the following point:
Here we can see that, in fact, Machine Learning has elements of both Data Science and Artificial Intelligence. To visualise this point:
Machine learning is the term used to describe the process by which a computer system produces insights from data and then using algorithms produces actions such as predicting outcomes, problem solving and learning from new information.
Let’s go back to the example of the travel agent’s weather prediction tool. To achieve the same output without Machine Learning we would use similar Data Science processes to analyse the data to produce the insights we need. But to look at the insights and create the predictions for future weather we would need human involvement or human intelligence rather. However with the use of Machine Learning algorithms we now give the computer system the ability to replace the human in this process i.e. we have added Artificial Intelligence.
Side Note revisited: Machine Learning is an application of Artificial Intelligence that DOES involves the use of data by computer systems.
We hope this bulletin helps clear things up between these three key terms used in the Data and Analytics industries. We love all things data and are proud to be a part of the Data and Analytics industry. If you are looking for a company that can make help with the recruitment needs of your business then feel free to get in touch.