Integrating Symmetry into Differentiable Planning with Steerable Convolutions
Abstract
To achieve this, we draw inspiration from equivariant convolution networks and model the path planning problem as a set of signals over grids. We demonstrate that value iteration can be treated as a linear equivariant operator, which is effectively a steerable convolution. Building upon Value Iteration Networks (VIN), we propose a new Symmetric Planning (SymPlan) framework that incorporates rotation and reflection symmetry using steerable convolution networks. We evaluate our approach on four tasks: 2D navigation, visual navigation, 2 degrees of freedom (2-DOF) configuration space manipulation, and 2-DOF workspace manipulation. Our experimental results show that our symmetric planning algorithms significantly improve training efficiency and generalization performance compared to non-equivariant baselines, including VINs and GPPN.
Cite
Text
Zhao et al. "Integrating Symmetry into Differentiable Planning with Steerable Convolutions." International Conference on Learning Representations, 2023.Markdown
[Zhao et al. "Integrating Symmetry into Differentiable Planning with Steerable Convolutions." International Conference on Learning Representations, 2023.](https://mlanthology.org/iclr/2023/zhao2023iclr-integrating/)BibTeX
@inproceedings{zhao2023iclr-integrating,
title = {{Integrating Symmetry into Differentiable Planning with Steerable Convolutions}},
author = {Zhao, Linfeng and Zhu, Xupeng and Kong, Lingzhi and Walters, Robin and Wong, Lawson L.S.},
booktitle = {International Conference on Learning Representations},
year = {2023},
url = {https://mlanthology.org/iclr/2023/zhao2023iclr-integrating/}
}