ASBI: Leveraging Informative Real-World Data for Active Black-Box Simulator Tuning
This addresses the challenge of parameter estimation in black-box simulators for robotics, offering a practical solution for tasks like object simulation, though it is incremental as it builds on existing simulation-based inference methods.
The paper tackles the problem of tuning black-box simulators in robotics by introducing Active Simulation-Based Inference (ASBI), which uses robots to actively collect real-world data to optimize parameters, achieving accurate parameter estimation with posteriors sharply concentrated around true values in simulations and a real robot application.
Black-box simulators are widely used in robotics, but optimizing their parameters remains challenging due to inaccessible likelihoods. Simulation-Based Inference (SBI) tackles this issue using simulation-driven approaches, estimating the posterior from offline real observations and forward simulations. However, in black-box scenarios, preparing observations that contain sufficient information for parameter estimation is difficult due to the unknown relationship between parameters and observations. In this work, we present Active Simulation-Based Inference (ASBI), a parameter estimation framework that uses robots to actively collect real-world online data to achieve accurate black-box simulator tuning. Our framework optimizes robot actions to collect informative observations by maximizing information gain, which is defined as the expected reduction in Shannon entropy between the posterior and the prior. While calculating information gain requires the likelihood, which is inaccessible in black-box simulators, our method solves this problem by leveraging Neural Posterior Estimation (NPE), which leverages a neural network to learn the posterior estimator. Three simulation experiments quantitatively verify that our method achieves accurate parameter estimation, with posteriors sharply concentrated around the true parameters. Moreover, we show a practical application using a real robot to estimate the simulation parameters of cubic particles corresponding to two real objects, beads and gravel, with a bucket pouring action.