I am a research associate (chercheur qualifié) of the fund for scientific research F.R.S.-FNRS of Belgium's French Community. I am affiliated with IRIDIA, CoDE, Université Libre de Bruxelles, Brussels, Belgium.
My research focuses mainly on computational intelligence and in particular on the automatic design of swarm intelligence systems and multi-robot systems.
My research interests include also metaheuristics and local search methods for combinatorial optimization. I am particularly interested in the application of statistical methods, design of experiments, and machine learning techniques, notably for assessing the performance of algorithms and for fine-tuning their parameters.
I am also interested in epistemology and in the philosophical foundations of computational intelligence.
I served as a member of the organizing committee for 15 international conferences, and as a member of the program committee for more than 50 international conferences in the last five years. Since 2002, I am a member of the organizing committee of the series of conferences ANTS.
Swarmanoid, the movie
The video "Swarmanoid, the movie" won the Best Video Award at the AAAI-11 AI Video Competition!
This video, written and directed by Mauro Birattari and Rehan O'Grady, presents the main results of the Swarmanoid project, a FET-OPEN project funded by the European Commission and coordinated by Marco Dorigo. For more information, see the website of the Swarmanoid project.
A machine learning perspective
M. Birattari. Springer, 2009
About this book
The importance of tuning metaheuristics is widely acknowledged. However, there is very little dedicated research on the subject. Typically, scientists and practitioners tune metaheuristics by hand, guided only by their experience and by some rules of thumb. Tuning metaheuristics is often considered to be more of an art than a science.
This book lays the foundations for a scientific approach to tuning metaheuristics. The fundamental intuition that underlies Birattari's approach is that the tuning problem has much in common with the problems that are typically faced in machine learning. By adopting a machine learning perspective, the author gives a formal definition of the tuning problem, develops a generic algorithm for tuning metaheuristics, and defines an appropriate experimental methodology for assessing the performance of metaheuristics.