The VUB Data Lab is a cross-disciplinary research lab at the Solvay Business School of the Vrije Universiteit Brussel (VUB) in Belgium. We focus on developing and adapting analytical tools and algorithms which retrieve valuable insights from data while taking into account the specific operational context, domain expertise, and business requirements of the application at hand. This approach differs from the typical statistical or IT perspective. Moreover, we exploit the connection between data analysis and the properties of complex systems. In doing so, we develop data analysis tools based on information theory, nonlinear dynamics, ergodic theory, and statistical physics.
At the VUB Data Lab, we harness the power of Large Language Models (LLMs) to advance natural language understanding and scientific progress, applying them to domains ranging from the science of science to practical implementations. Our work also includes developing Explainable AI methods to uncover and address algorithmic bias, leveraging Graph Neural Networks to analyze structured data, and tackling the challenges of multitask learning with inspiration from physics and deep learning theory. We model biological systems using statistical physics, study wealth inequality through computational approaches, and apply tools like ergodic theory and nonlinear dynamics to explore the evolution of complex systems. Through these diverse research topics, we bridge theoretical innovation and real-world impact, addressing pressing challenges across science, business, and society.
The lab is headed by prof. Vincent Ginis, prof. Sam Verboven, prof. Filip Van Droogenbroeck, and Prof. Marie-Anne Guerry. We have strong ties with the MOBI research center on Mobility, Logistics, and Electric vehicles, with the AI research center, doing interdisciplinary research on Artificial Intelligence, as well as with the Applied Physics research group. The lab performs both fundamental and applied research, always in close collaboration with a wide network of (inter)national academic and industrial partners.