Bluffer’s guide to AI, Deep Learning and Data Science

The wonderful thing about buzzwords like “AI” is that they capture a movement, the mood of the times. While AI’s predecessor “big data” sounded macho and kind of sinister, “artificial intelligence” is more relatable, appearing amiably in our lives in the guise of Siri and Alexa. The idea of AI keys strongly into our hopes and fears for a future where we share space with machines and robots.

The terrible thing about buzzwords is that when you come to think about them rationally in the context of your business, they’re far too imprecise. The data world already suffers from an explosion of concepts and buzzwords, and adding AI to the mix doesn’t help much.

So I thought I’d make a little chart to show the relationship between data science, machine learning, artificial intelligence, and its poster child, deep learning.

Having laid out the relationships, I’d like to offer my high-level definitions of each area. As always, there are debatable boundaries, and much more detail that I could dive into, but this should be good enough to orient a business conversation.

  • Artificial intelligence: a specialism of computer science, focused on giving computers capabilities that imitate aspects of human intelligence. Such areas include pattern recognition, computer vision, and reasoning.
  • Machine learning: a class of statistical methods of AI, which enables a machine to classify inputs or make predictions about future behavior, based on having been trained with similar previous data. Applications include spam detection, sales forecasting, product recommendations. Common techniques include linear and logistic regression, decision trees, k-means clustering, naive Bayes, support vector machines.
  • Deep learning: a specialization of one particular machine learning technique, artificial neural networks, that enables machines to train themselves given sufficiently large amounts of example data. Applications include image and speech recognition, language translation, autonomous vehicles. Deep learning techniques vary, with specializations for particular applications.
  • Data science: an interdisciplinary field of statistics, computer science and business skills that uses data to understand and affect behavior of people, systems and environments.

If you liked this and want to hear more, please join my newsletter.