Equivariant Deep Dynamical Model for Motion Prediction

Abstract

Learning representations through deep generative modeling is a powerful approach for dynamical modeling to discover the most simplified and compressed underlying description of the data, to then use it for other tasks such as prediction. Most learning tasks have intrinsic symmetries, i.e., the input transformations leave the output unchanged, or the output undergoes a similar transformation. The learning process is, however, usually uninformed of these symmetries. Therefore, the learned representations for individually transformed inputs may not be meaningfully related. In this paper, we propose an SO(3) equivariant deep dynamical model (EqDDM) for motion prediction that learns a structured representation of the input space in the sense that the embedding varies with symmetry transformations. EqDDM is equipped with equivariant networks to parameterize the state-space emission and transition models. We demonstrate the superior predictive performance of the proposed model on various motion data.

Accepted at AISTATS 2022

References in corpus (7)

• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• A Disentangled Recognition and Nonlinear Dynamics Model for Unsupervised Learning
• E(n) Equivariant Graph Neural Networks
• Permutation-equivariant neural networks applied to dynamics prediction
• Recurrent Kalman Networks: Factorized Inference in High-Dimensional Deep Feature Spaces
• Deep Switching Auto-Regressive Factorization:Application to Time Series Forecasting
• Circular-Symmetric Correlation Layer based on FFT