Humboldt-Universität zu Berlin - School of Business and Economics

The Chair of AI in Business and Economics explores how artificial intelligence and data science can advance research and decision-making in economics and the social sciences. Our goal is to bridge the gap between modern machine learning/AI methods and robust statistical principles, developing transparent, interpretable, and trustworthy tools for applied research. We work across disciplines and institutions, combining theory, computation, and policy relevance to address some of today’s most pressing economic and societal challenges.

Katarzyna's research

Katarzyna Reluga’s research lies at the intersection of statistical machine learning, causal inference, and explainable AI. Her work focuses on creating methods that enhance the interpretability, robustness, and fairness of AI-driven models, with a strong emphasis on applications in public policy and official statistics. She integrates modern AI techniques with classical frameworks such as survey sampling and small-area estimation to improve the precision and transparency of socioeconomic indicators.

Building on her background in post-selection and computational inference, she develops methodologies that connect theory with practice, addressing real-world problems in areas including public health, poverty mapping, and policymaking. Her research aims to advance statistical methodology and data science at the interface between statistics and AI, promoting open, reproducible, and socially responsible research.

For a list of current projects, see my website. 

For questions about writing a Bachelor/Master/PhD thesis with me, see Theses section.

Philipp's research

Philipp Warode's research lies at the intersection of algorithmic game theory, optimization, and network design, with a focus on computational and algorithmic aspects of equilibria and flow problems in networks. He develops mathematical and algorithmic tools to understand how equilibria behave under parametric changes, such as varying demands or pricing schemes, and to design efficient algorithms for computing them. This includes work on congestion and Wardrop equilibria, carbon pricing in traffic networks, and parametric flow problems with nonlinear or convex costs.

Philipp is also interested in approximation algorithms and scheduling theory, particularly in developing simple and provably efficient strategies for complex or uncertain environments such as non-clairvoyant scheduling. Broadly, his research combines combinatorial optimization, convex analysis, and algorithmic game theory to address fundamental questions in computational efficiency and equilibrium design.