Interactive Differentiable Simulation

TL;DR

Interactive Differentiable Simulation (IDS) is a physics engine that enables interpretable, accurate inference of rigid-body dynamics from visual data, significantly improving sample efficiency in control tasks.

Contribution

We introduce IDS, a differentiable physics engine that integrates with deep learning for interpretable system identification and control, surpassing prior methods in efficiency.

Findings

IDS enables accurate physical property inference from visual input.

Integration with control algorithms improves sample efficiency by orders of magnitude.

Demonstrates effectiveness in nonlinear dynamical system tasks.

Abstract

Intelligent agents need a physical understanding of the world to predict the impact of their actions in the future. While learning-based models of the environment dynamics have contributed to significant improvements in sample efficiency compared to model-free reinforcement learning algorithms, they typically fail to generalize to system states beyond the training data, while often grounding their predictions on non-interpretable latent variables. We introduce Interactive Differentiable Simulation (IDS), a differentiable physics engine, that allows for efficient, accurate inference of physical properties of rigid-body systems. Integrated into deep learning architectures, our model is able to accomplish system identification using visual input, leading to an interpretable model of the world whose parameters have physical meaning. We present experiments showing automatic task-based robot design and parameter estimation for nonlinear dynamical systems by automatically calculating gradients in IDS. When integrated into an adaptive model-predictive control algorithm, our approach exhibits orders of magnitude improvements in sample efficiency over model-free reinforcement learning algorithms on challenging nonlinear control domains.