Tea Time Talks 2024: Bounding-Box Inference for Error-Aware Model-Based RL - Erin Talvitie