Data intelligence concerns how data may be acquired, handled, managed, enriched, processed in order to extract, or to reveal information, ultimately to make decisions or serve as a decision aid. Here, data should be understood in a very broad sense, originating from various sources (sensors, databases, social networks, etc), involving various types (images, texts in natural language, numbers, etc). In this context, data has to be organized; vast amounts of data have to be dealt with; data is usually streamed, and its characteristics may change over time; useful information has to be extracted from this flow of data. The fields of applications are extremely large: business related applications are well-known, but there are also a wide range of applications in science (physics, astronomy, biology, medicine, …).

In CRIStAL, the “Data Intelligence Group” (DatInG) is composed of a set of 3 research groups: MAGNET, SCOOL and SIGMA. DatInG combines strong skills in machine learning, data mining, and signal processing. LINKS considers graph databases for storing and querying information networks. MAGNET considers trustworthy machine learning (privacy, fairness, decentralized/federated learning) and Natural Language Processing. SIGMA considers signals, i.e. mathematical entities coming from physical sensors. SCOOL investigates sequential decision making under uncertainty problems, mostly bandit and reinforcement learning problems.

Altogether, these 3 groups address the most important steps of the data intelligence pipeline. Activities in DatInG span from fundamental research to applications and collaborations with companies.