Equivariant Neural Operator Learning with Graphon Convolution

Authors: Chaoran Cheng, Jian Peng

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through extensive experiments on large-scale electron density datasets, we observed that our model significantly outperformed the current state-of-the-art architectures. Multiple ablation studies were also carried out to demonstrate the effectiveness of the proposed architecture.
Researcher Affiliation Academia Chaoran Cheng University of Illinois Urbana-Champaign chaoran7@illinois.edu Jian Peng University of Illinois Urbana-Champaign jianpeng@illinois.edu
Pseudocode No The paper describes its methods but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks.
Open Source Code Yes Our code is publicly available at https://github.com/ccr-cheng/Inf GCN-pytorch.
Open Datasets Yes QM9. The QM9 dataset [41, 39] contains 133,885 species with up to nine heavy atoms (CONF). The density data as well as the data split come from [17, 18]... Cubic. This large-scale dataset contains electron densities on 17,418 cubic inorganic materials [53]... MD. The dataset contains 6 small molecules... The former 4 molecules are from [1]... The latter two are from [2].
Dataset Splits Yes QM9... gives 123835 training samples, 50 validation samples, and 10000 testing samples.
Hardware Specification Yes All models were trained on a single NVIDIA A100 GPU.
Software Dependencies No Our implementation was based on the e3nn3 package which implements efficient spherical vector manipulations. No specific version numbers for software dependencies are provided.
Experiment Setup Yes We set the maximal degree of spherical tensors to L = 7, with 16 radial basis and 3 convolution layers. [...] The training and testing specifications are provided in Table 4. The Table includes details such as 'cutoff', 'n_iter', 'lr', 'patience', 'batch_size', 'lr_decay', 'train_sample', 'inf_sample' for different models and datasets.