Data Science
Data scientists work with huge volumes of information, create predictions, and extract insights. For example in commerce: a customer looking at a fishing rod and hook will automatically receive an offer for line and complementary products that other customers purchased.
In e-commerce, Data Science is used to predict demand and supply. Graphics card manufacturers know when new games will be released and what the requirements are, but it's difficult to estimate how many GPUs gamers will buy. Multi-year data analysis using Data Science allows estimating the number.
Data Mining
Data Mining processes information and extracts useful knowledge from it that is organized in convenient structures. Classification is based on hidden patterns and anomalies. The technology is also common in the business world.
In business, data mining is used to analyze shopping carts to discover which products are bought together and solve additional tasks. In e-commerce, product sets are defined according to complex patterns with dozens of attributes.
Machine Learning
Today, humans still program machines. The goal is to free them from the role of "teacher" – to allow the machine to learn automatically from massive data.
We encounter this daily: spam filtering is based on machine learning. The robot identifies spam characteristics and sends the email to the appropriate folder; it also needs positive examples of regular emails. Most filters work well – check your spam folder.
In this case, it's supervised learning: the robot learns based on a dataset with spam characteristics and emails marked by users. It analyzes new information and expands its knowledge.
Deep Learning
This is a subfield of machine learning based on biologically inspired neural networks, enabling work with even larger volumes of data.
Think about software that recognizes images: for a machine, this is a very complex task. To distinguish between a bus and a loaf of bread, especially when colors are similar, it needs to be trained on thousands of uniformly sized images to accumulate enough knowledge.
Deep learning is used in a huge variety of applications: speech-to-text recognition, drug discovery in pharma, Google Translate, and more. Neural networks have raised the level of machine learning and enabled solving much more difficult tasks.
The Evolution of Artificial Intelligence
When talking about AI, many imagine cinematic "super-intelligence". In the business world, it refers to a collection of software solutions that perform a series of tasks without human involvement – planning, inferring, and deciding based on existing data.
We're still in the era of narrow AI, which only does what it was taught, but progress is rapid: Data Science, Data Mining, deep learning, and machine learning are making it smarter.
Autonomous cars, industrial robots, virtual assistants, chatbots, speech recognition, and more – all indicate that many scenes from science fiction movies are not far off. It's possible that within a few decades we'll see AI at the level of human consciousness, but already today there are AI-based solutions that advance businesses.
