Gaussian Plane-Wave Neural Operator for Electron Density Estimation
Authors: Seongsu Kim, Sungsoo Ahn
ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on QM9, MD, and material project datasets demonstrate GPWNO s superior performance over ten baselines. |
| Researcher Affiliation | Academia | 1Pohang University of Science and Technology (POSTECH), Pohang, South Korea. Correspondence to: Seongsu Kim <seongsukim@postech.ac.kr>, Sungsoo Ahn <sungsoo.ahn@postech.ac.kr>. |
| Pseudocode | No | The paper describes its methods verbally and through diagrams but does not include any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our implementation of the GPWNO is published in the Git Hub3. 3https://github.com/seongsukim-ml/GPWNO |
| Open Datasets | Yes | We empirically verify the effectiveness of our methods on three datasets: QM9 (Ruddigkeit et al., 2012; Ramakrishnan et al., 2014; Jørgensen & Bhowmik, 2022), MD (Brockherde et al., 2017; Bogojeski et al., 2020) datasets. We additionally benchmark our algorithm on the newly curated datasets from the material project (Jain et al., 2013; Shen et al., 2021, MP). |
| Dataset Splits | Yes | train/val/test split 123835/50/1600 (Table 4, QM9) [...] The size of the datasets ranges from 7,681 to 26,081. We use 1,000 or 500 samples for the test and validation, depending on the dataset size. |
| Hardware Specification | No | The paper does not specify the exact hardware (e.g., GPU models, CPU types, or cloud instance specifications) used for running the experiments. |
| Software Dependencies | No | The paper mentions the use of the ADAM optimizer and refers to VASP and Quantum ESPRESSO for dataset generation, but it does not provide specific version numbers for any key software libraries or frameworks (e.g., Python, PyTorch, TensorFlow) used for its model implementation. |
| Experiment Setup | Yes | We list the training specifications and major hyperparameters of our experiments in Table 7. [...] Key hyperparameters in our model include the spherical harmonics (sp) order in the GTO layer and the radius cutoff distance for each message-passing operation. [...] Additionally, the probe cutoff (p.cutoff) and query cutoff (q.cutoff) parameters specify the radius distances used in Equation (10) and Equation (13), respectively. |