KDD: This was in fact the first term ever used to describe what we are still actively pursuing today: knowledge discovery from data(bases). Initially, this was the basic function assigned to what we call AI today: to get information out of databases by combining and comparing data, and to use this information to obtain new insights/knowledge. Actually, KDD goes all the way back to when SAS was originally founded in 1976!
Data mining: This term refers to a next level of data discovery: the purpose-oriented search for meaningful patterns in data. Churn detection (analyzing customers’ behaviour in regard to the competition) and association rule mining (studying grouped purchases of products) are two of the more widely known applications of data mining.
Analytics: As the word suggests, analytics goes beyond a mere search for correlations: it commonly refers to the discovery, interpretation, and communication of meaningful patterns in data and often also to the process of applying those patterns towards effective decision making. Data mining is therefore often an integral part of the process, but the link between data and decision making becomes more apparent here.
Machine learning and deep learning: These two terms are often used interchangeably, with deep learning typically defined as the most advanced form. Both refer to a form of artificial intelligence where the machine automatically starts learning patterns from data whereby the results and the feedback provided by humans are used to continuously improve itself. Thus, the results gradually get more refined and precise and the value for business applications grows incrementally.
Data science: Can be compared to data mining, but encompasses a profound featurization strategy where variables that have a big impact on the predictions are searched for. The term data science is aptly coined: it elevates the art of analytics to academic level on the one hand, and it gives birth to the ‘data scientist’, the sexiest job around because it manages to combine everything: business ànd technology ànd science/academics.
Decision science: I wanted to add this final term because it reflects my prediction for this market. Decision science would then be more or less the same technology as before, but with the focus shifting from data and research to business and value. The user interface will be adapted to better support decision making processes, but the underlying technology will still be as intelligent and powerful as ever before, and even more.
AI: The term ‘AI’ has been around for quite a while now. But today it refers to more application fields than it did at first. We use the term ‘Artificial intelligence (AI)’ whenever technology is involved that makes it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks. Most AI examples that you hear about today – from chess-playing computers to self-driving cars – rely heavily on deep learning and natural language processing.
You may have noticed that there is lots of overlap between the various definitions. And that’s no surprise, because one technology often organically evolves from the other. Essentially, decision science could therefore still be called AI, according to the above definitions. But by coining and understanding the various terms, we can better understand at which level a specific solution should be situated (or at which level the vendor wants you to situate it).
But in the end, the most important lesson to learn is: technology becomes more powerful and more intelligent each day, and we can derive ever more business value from it, provided we have the right talent available to create, use and interpret the available intelligence.