Anna M. Maddux

PhD student in the [Sycamore Lab](https://www.epfl.ch/labs/sycamore/). Laboratoire Automatique, Institute of Mechanical Engineering, EPFL, Switzerland.

20240311_skitouring2_cropped.png

EPFL STI IGM SYCAMORE

Office: ME C1 402

1015 Lausanne, Switzerland

I am currently enrolled in a PhD program at EPFL Lausanne, where I am conducting research under the supervision of Maryam Kamgarpour in the Systems Control and Multiagent Optimization Research (Sycamore) Lab. My research interests lie in the fields of game theory, mechanism design and learning in multi-agent systmes. In my research, I develope learning algorithms with convergence- and sample complexity guarantees for repeated games. In particular, I focus on games with unknown constraints for which I establish learning algorithms that result in an feasible outcome. I am also very interested in understaning to which outcome differenet learning algorithms converge and whether an efficient equilibrium in terms of social welfare can be attained by different learning algorithms. To validate my learning algorithms I consider various application areas ranging from shared mobility systems, energy markets, and multi-robot systems.

Prior to my PhD, I obtained a MSc. in Statistics from ETH Zürich (2021), where I did my master’s thesis at the Automatic Control Laboratory (Ifa) under the supervision of Florian Dörfler and Fadoua Balabdaoui. I, furthermore, obtained a BSc. in Mathematics (2019) and a BSc. in Economics (2018) from the Albert-Ludwigs-University Freiburg. In the past, I interned at Daimler Benz and the German federal parliament. Throughout my undergraduate and graduate studies I have been a fellow of the Friedrich Nauman Foundation. I was also a fellow of the Accenture female talent program.

Links:      Email     Google Scholar     ORCID     GitHub     LinkedIn

Misc: My Erdos number is 5.

Announcements

Jul, 2024 I am presenting my work on at Finite-time convergence to an ε-efficient Nash equilibrium in potential games at the Workshop on Learning in Games taking place in Metz.
Jun, 2024 I am excited to be attending the Games and Artificial Intelligence Multidisciplinary Summer School summer school (GAIMSS24) taking place in Metz.
May, 2024 I am presenting my poster on Multi-Agent Learning in Contextual Games under Unknown Constraints at AISTATS taking place in Valencia.
Apr, 2024 I am a co-organizer of the NCCR Automation Junior Retreat taking place April, 15 to 17.
Feb, 2024 I was invited to the Kantonsschule Beromünster to present my research under the NCCRWomen campaign on February, 23.

Selected Publications

  1. Data-driven behaviour estimation in parametric games
    Anna M Maddux ,  Nicolò Pagan ,  Giuseppe Belgioioso ,  and  Florian Dörfler
    IFAC-PapersOnLine, 2023
  2. Multi-Agent Learning in Contextual Games under Unknown Constraints
    Anna M Maddux ,  and  Maryam Kamgarpour
    International Conference on Artificial Intelligence and Statistics (AISTATS), 2024
  3. Finite-time convergence to an ε-efficient Nash equilibrium in potential games
    Anna Maddux ,  Reda Ouhamma ,  and  Maryam Kamgarpour
    ICML Workshop on Foundations of Reinforcement Learning and Control: Connections and Perspectives, 2024, 2024