Abstract: Image convolutions have been a cornerstone of a great number of deep learning advances in computer vision. The research community is yet to settle on an equivalent operator for sparse, unstructured continuous data like point clouds and event streams however. We present an elegant sparse matrix-based interpretation of the convolution operator for these cases, which is consistent with the mathematical definition of convolution and efficient during training. On benchmark point cloud classification problems we demonstrate networks built with these operations can train an order of magnitude or more faster than top existing methods, whilst maintaining comparable accuracy and requiring a tiny fraction of the memory. We also apply our operator to event stream processing, achieving state-of-the-art results on multiple tasks with streams of hundreds of thousands of events.

SlidesLive

Similar Papers

FKAConv: Feature-Kernel Alignment for Point Cloud Convolution
Alexandre Boulch (valeo.ai)*, Gilles Puy (Valeo), Renaud Marlet (Ecole des Ponts ParisTech)
Quantum Robust Fitting
Tat-Jun Chin (University of Adelaide), David Suter (Edith Cowan University), Shin-Fang Ch'ng (The University of Adelaide)*, James Quach (The University of Adelaide)
Progressive Batching for Efficient Non-linear Least Squares
Huu Le (Chalmers University of Technology)*, Christopher Zach (Chalmers University), Edward Rosten (Snap Inc.), Oliver J. Woodford (Snap Inc)