← Search

George Matheos

2 accepted papers

2024

Partially Observable Task and Motion Planning with Uncertainty and Risk Awareness

RSS 2024poster

Integrated task and motion planning (TAMP) has proven to be a valuable approach to generalizable long-horizon robotic manipulation and navigation problems. However, the typical TAMP problem formulation assumes full observability and deterministic action effects. These assumptions limit the ability o…

Cited by 11SourcePDFScholar
2023

SMCP3: Sequential Monte Carlo with Probabilistic Program Proposals

AISTATS 2023poster

This paper introduces SMCP3, a method for automatically implementing custom sequential Monte Carlo samplers for inference in probabilistic programs. Unlike particle filters and resample-move SMC (Gilks and Berzuini, 2001), SMCP3 algorithms can improve the quality of samples and weights using pairs o…