Embedded Files
Real-world uncertainty in any decision-making problem does not come with an exact and precise description of the noise distribution. This could be due to a noise distribution that is estimated using limited data, changes in the underlying environment, or rare but significant catastrophic events. The standard approach of stochastic optimization avoids all such concerns by assuming that the underlying noise distribution is fully known and remains constant during optimization and, thus, its policies tend to exhibit poor performance in the first moment this assumption becomes invalid.
In this context, distributionally robust optimization provides a framework to develop decisions which remain robust even in the case when the exact description of the underlying distribution is not known and it belongs to some ambiguity set of distributions. During recent years, DRO became a popular technique which was adopted to robust control design, reinforcement learning, the convergence analysis of optimization algorithms, and even some equilibrium problems with distributional uncertainty.
This workshop brings together people working on these threads from a few different angles:
distributionally robust control and decision-making, including robust policy synthesis and safety guarantees,
DRO applied to learning algorithms and their analysis,
game-theoretic formulations of distributional robustness
applications ranging from robot navigation to carbon-aware computing.
Speakers include researchers from Princeton, Cornell, UC San Diego, TU Delft, Imperial College London, the University of Chicago Booth School, ETH Zürich, and the University of Salerno.
Attendees
Attendees
This workshop is designed for anyone interested in using techniques from distributionally robust optimization in the analysis and design of feedback control loops, learning algorithms, and game-theoretic dynamics.
This workshop is designed for anyone interested in using techniques from distributionally robust optimization in the analysis and design of feedback control loops, learning algorithms, and game-theoretic dynamics.
We particularly welcome participation from early-career researchers, PhD students and postdocs, as well as professionals from industry.
We particularly welcome participation from early-career researchers, PhD students and postdocs, as well as professionals from industry.
We encourage applications from universities in geographical regions that are under-represented in the control community. Information about the CSS travel grant can be found here.
We encourage applications from universities in geographical regions that are under-represented in the control community. Information about the CSS travel grant can be found here.
Speakers
Speakers
Nikolay Atanasov
UC San Diego
Sergio Grammatico
TU Delft
Sophie Hall
ETH Zürich
Dario Paccagnan
Imperial College London
Giovanni Russo
University of Salerno
Bahar Tașkesen
University of Chicago Booth School of Business
Soroosh Shafiee
Cornell University
Bartolomeo Stellato
Princeton University
Organizers
Riccardo Cescon, EPF Lausanne
Riccardo Cescon, EPF Lausanne
Riccardo Cescon is a PhD candidate in the Automatic Control Laboratory (LA) at EPF Lausanne. He received the B.Sc. degree in information engineering in 2020, and the M.Sc. degree in automation engineering in 2022, both from the University of Padova. His research interests include distributionally robust optimization with applications in control theory and machine learning.
Nicolas Lanzetti, California Institute of Technology
Nicolas Lanzetti, California Institute of Technology
Nicolas Lanzetti is a postdoctoral scholar in the Department of Computing and Mathematical Sciences at the California Institute of Technology, hosted by Profs. Adam Wierman, Eric Mazumdar, and Steven Low. He completed his PhD at the Automatic Control Laboratory at ETH Zurich, under the supervision of Prof. Florian Dörfler (main advisor) and Prof. Alessio Figalli (second advisor). Previously, he received his B.Sc. and M.Sc. in Mechanical Engineering, with a focus on Robotics, Systems, and Control from ETH Zurich in 2016 and 2019, respectively. He was a visiting researcher at the Massachusetts Institute of Technology, Stanford University, and EPFL. His research enables AI-driven decision-making in critical engineering, business, and societal infrastructure by developing the theoretical and algorithmic foundations for strategic dynamic decision-making under uncertainty.
Andrea Martin, KTH Royal Institute of Technology
Andrea Martin, KTH Royal Institute of Technology
Andrea Martin is a Postdoctoral Researcher in the Department of Decision and Control Systems at KTH Royal Institute of Technology. He received his Ph.D. in Robotics, Control, and Intelligent Systems from EPFL in 2025. Before that, he obtained a B.Sc. in Information Engineering and two M.Sc. degrees in Automation Engineering and Automatic Control and Robotics from the University of Padova and the Polytechnic University of Catalonia through the TIME double degree program. His research interests lie at the intersection of control theory, optimization, and machine learning. He was awarded the Digital Futures Postdoctoral Fellowship in 2024 and the Swiss National Science Foundation Postdoc.Mobility Fellowship in 2025.
Google Sites
Report abuse