Clebsch–Gordan Nets: a Fully Fourier Space Spherical Convolutional Neural Network

Authors: Risi Kondor, Zhen Lin, Shubhendu Trivedi

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section we describe experiments that give a direct comparison with those reported by Cohen et al. [1]. We choose these experiments as the Spherical CNN proposed in [1] is the only direct competition to our method. Rotated MNIST on the Sphere...Atomization Energy Prediction Next, we apply our framework to the QM7 dataset...3D Shape Recognition Finally, we report results for shape classification using the SHREC17 dataset [32]...Table 1: Results on the QM7 and 3D shape recognition datasets.
Researcher Affiliation Academia Risi Kondor1 Zhen Lin1 Shubhendu Trivedi2 1The University of Chicago 2Toyota Technological Institute {risi, zlin7}@uchicago.edu, shubhendu@ttic.edu
Pseudocode Yes The algorithm is presented in explicit form in the Supplement.
Open Source Code No While the CG transform is not currently available as a differentiable operator in any of the major deep learning software frameworks, we have developed and will publicly release a C++ Py Torch extension for it.
Open Datasets Yes Next, we apply our framework to the QM7 dataset [30, 31]... Finally, we report results for shape classification using the SHREC17 dataset [32], which is a subset of the larger Shape Net dataset [33]...
Dataset Splits No We report three sets of experiments: For the first set both the training and test sets were not rotated (denoted NR/NR), for the second, the training set was not rotated while the test was randomly rotated (NR/R) and finally when both the training and test sets were rotated (denoted R/R).
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU models, or cloud computing instance types used for running the experiments.
Software Dependencies No While the CG transform is not currently available as a differentiable operator in any of the major deep learning software frameworks, we have developed and will publicly release a C++ Py Torch extension for it.
Experiment Setup No more details about the data, baseline models, as well as the detailed architecture of our model and hyperparameters are provided in the appendix.