Statement Robert Mateescu

My research interests are in artificial intelligence. I focus primarily on automated reasoning in graphical models, such as Bayesian networks, constraint networks, Markov networks or influence diagrams, which have become a central paradigm for knowledge representation and reasoning. My goal is to develop efficient algorithms and data structures for such models, both on a theoretical and practical level. The guiding principle of my approach is the exploitation of the structural information revealed by the underlying graphical representation. There are numerous examples of problems defined as graphical models, including design, scheduling, planning, diagnosis, decision making or genetic linkage analysis. Solving such problems is computationally hard, and typically requires handling uncertainty (probabilistic information) or inconsistency (deterministic information). Advances in exact or approximate methods are therefore crucial, with potential impact across many disciplines. I believe that many new improvements are still possible, through an interdisciplinary approach, combining state-of-the-art algorithms and insights from graph theory, probability theory, statistics, statistical physics, artificial intelligence, logic, formal verification or information theory, to name only those which have already played a part in my research so far, as detailed further.