Keywords: water-pouring, probabilistic model
Abstract: This report presents a probabilistic model designed to replicate human performance in water-pouring tasks. Leveraging Bayesian inference and a probabilistic intuitive physics engine, the model can predict pouring angles through imagined actions using noisy observations. The model aligns closely with human behavior, showcasing its ability to tackle noise in simulated doing, and to handle mental simulations
without concurrent motor activity
Supplementary Material: zip
Submission Number: 233
Loading