Abstract:
Optimization is the bedrock of robotics, and robotics is a great stress
test for optimization methods. High dimensionality, nonconvexity, and
nonsmoothness make many robotics problems notoriously hard to solve,
whether with learning-based or model-based strategies.
In this talk, I will discuss our recent advances in bringing
learning-based techniques together with optimization principles to
address this. First, I will present extensions of moment-based
optimization to data-driven, nonparametric problems that retain partial
global optimality guarantees. I will discuss how these methods enable
new approaches to high-dimensional, contact-rich manipulation. Next, I
will explore how foundation models can be combined with global
optimization to reduce modeling burdens and improve generalization of
model-based solvers. I will conclude with an outlook on using
optimization principles to reshape latent geometry for world-model
learning. Together, these examples illustrate how learning can expand
the reach of optimization, while optimization can bring structure and
guarantees to data-driven robotics.

Bio:
Frederike Dümbgen is an assistant professor in the Department of
Mechanical Engineering at Carnegie Mellon University, where she leads
the Trustworthy Robotics, Intelligence, and Optimization (TRIO) Lab. Her
research combines optimization and artificial intelligence to develop
scalable and trustworthy robotic systems, with a particular emphasis on
global optimization and optimality guarantees.
Before joining Carnegie Mellon, she was a researcher in the WILLOW team
at Inria Paris and a postdoctoral fellow at the Robotics Institute of
the University of Toronto. She holds a Ph.D. in Computer and
Communication Sciences from École polytechnique fédérale de Lausanne
(EPFL), Switzerland, where she also earned her B.Sc. and M.Sc. in
Mechanical Engineering. She was an RSS Pioneer in 2024 and received a
Google Women Techmakers Scholarship in 2020. She serves as a co-chair of
the IEEE Robotics and Automation Society’s Technical Committee on
Optimization for Robotics.