Maegan Tucker

Synthesizing optimal queries in a continuous space slashes the computational cost of active preference learning, enabling faster and more efficient preference elicitation.

Guaranteeing robustness for legged robots just got easier: this work provides a computationally tractable method to compute invariant sets around periodic orbits, enabling controller design that explicitly maximizes stability margins.

Guaranteeing robustness of hybrid systems just got easier: this work offers a computationally tractable method to find invariant sets with probabilistic guarantees, even when dynamics are only accessible through simulation.

Forget waiting hours: this MORL framework achieves 270x speedups on robotics tasks thanks to GPU-native parallelization.