Scaling Spherical CNNs

TL;DR

This paper presents methods to scale spherical CNNs for larger problems, achieving state-of-the-art results on molecular benchmarks and competitive weather forecasting performance by optimizing computations and input representations.

Contribution

The paper introduces novel model components, hardware-efficient implementations, and input representations to enable scaling of spherical CNNs to larger, more complex problems.

Findings

Achieved state-of-the-art on QM9 molecular benchmark

Demonstrated competitive weather forecasting results

Improved computational efficiency for spherical convolutions

Abstract

Spherical CNNs generalize CNNs to functions on the sphere, by using spherical convolutions as the main linear operation. The most accurate and efficient way to compute spherical convolutions is in the spectral domain (via the convolution theorem), which is still costlier than the usual planar convolutions. For this reason, applications of spherical CNNs have so far been limited to small problems that can be approached with low model capacity. In this work, we show how spherical CNNs can be scaled for much larger problems. To achieve this, we make critical improvements including novel variants of common model components, an implementation of core operations to exploit hardware accelerator characteristics, and application-specific input representations that exploit the properties of our model. Experiments show our larger spherical CNNs reach state-of-the-art on several targets of the QM9 molecular benchmark, which was previously dominated by equivariant graph neural networks, and achieve competitive performance on multiple weather forecasting tasks. Our code is available at https://github.com/google-research/spherical-cnn.