Research

Data, in all its forms and shapes, has been at the core of the Adrem Data Lab since its inception in 1979. Back then, the lab’s primary focus was on database management. Data mining and knowledge discovery followed soon after, bioinformatics and data-driven life sciences complemented our lab more recently. 

The importance of managing and understanding data has grown tremendously. All aspects of society now depend in some way on data. The amount and variety of data and the diversity of its use, resulted in an ever-changing landscape. It forms a fertile ground for interesting and diverse research questions, both from a foundational and practical point of view. The central mission of the Adrem data lab is to continue to play a leading and trend-setting role in the scientific research area of data science.  We further aim to capitalise on our expertise to expand our research in AI,  with the focus on AI-assisted data science.

Adrem Data Lab is a  key player in data mining, biomedical informatics, database theory and data management. We cover a plentitude of emerging problems in the data science field, such as:

• reasoning about uncertain, incomplete or missing data,
• managing semi-structured, schema-less, sequential, temporal and graph data,
• domain expert-driven feature engineering,
• building (explainable and -fair) predictive models,
• exploratory data analysis,
• modeling and analysing complex business processes,
• analysing and developing query languages for data analytics,
• and the study of frameworks for data cleaning.

The emergence of bioinformatics and other data-driven approaches to life sciences paved the way for translating and studying data mining and machine learning methods in the context of practical real-world problems, resulting in many life science collaborations. This has resulted in a high impact on diverse life science domains ranging from infectious diseases, oncology, toxicology, immunology and green biotechnology, often in collaboration with healthcare players such as hospitals and the pharma industry. Similar high impact holds for recent research in recommendation systems and its applications in e-commerce, with the successful Froomle spin-off as one of its results.

In all these areas high quality scientific results were obtained and research projects with collaborators (academic, public sector and industry) were established. At the same time, our approach to research remains largely fundamental in the sense that we aim at a precise mathematical formulation of research problems that allow not only for proposing solutions based on existing or new algorithms, but also for studying their intrinsic theoretical properties (complexity, expressiveness, optimality,..). Finally, the Adrem data lab is proud to have initiated certain important research directions in the areas of databases (scale independence, constraint-based data repairing), data mining (interactive pattern exploration, fairness), recommendation systems (recommendation framework for implicit feedback data) and life sciences (bioanalytical pattern discovery).