SE(3)-equivariant prediction of molecular wavefunctions and electronic densities

Authors: Oliver Unke, Mihail Bogojeski, Michael Gastegger, Mario Geiger, Tess Smidt, Klaus-Robert Müller

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To assess the ability of Phi SNet to predict molecular wavefunctions and electronic densities (see Fig. 1C), we train it on Kohn-Sham (...) for various non-equilibrium configurations of water, ethanol, malondialdehyde, uracil, and aspirin (...). The results are summarized in Tab. 5 and compared to the current state-of-the-art given by Sch NOrb [33].
Researcher Affiliation Collaboration 1 Machine Learning Group, Technische Universität Berlin, 10587 Berlin, Germany 2 DFG Cluster of Excellence Unifying Systems in Catalysis (Uni Sys Cat), Technische Universität Berlin, 10623 Berlin, Germany 3 Institute of Physics, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland 4 Computational Research Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720 5 Center for Advanced Mathematics for Energy Research Applications (CAMERA), Lawrence Berkeley National Laboratory, Berkeley, CA 94720 6 Department of Artificial Intelligence, Korea University, Anam-dong, Seongbuk-gu, Seoul 02841, Korea 7 Max Planck Institute for Informatics, Stuhlsatzenhausweg, 66123 Saarbrücken, Germany 8 BIFOLD Berlin Institute for the Foundations of Learning and Data, Berlin, Germany 9 BASLEARN TU Berlin/BASF Joint Lab for Machine Learning, Technische Universität Berlin, 10587 Berlin, Germany 10 Google Research, Brain Team, Berlin, Germany
Pseudocode No The paper describes the architecture and its building blocks in detail, and includes diagrams such as Figure 1B, but it does not provide a formal pseudocode block or algorithm listing.
Open Source Code No The paper discusses general concepts of open-source code and mentions 'open_source_code' as a potential benefit but does not explicitly state that the code for this paper's methodology is open-source or provide a link.
Open Datasets Yes Datasets for all molecules are taken from [33], with the exception of aspirin, for which geometries were sampled from the MD17 dataset [5] (more details on the datasets, training procedure, and hyperparameter settings can be found in Sections F and G of the supplement).
Dataset Splits No The paper mentions 'training procedure' in Section 5 and refers to 'Sections F and G of the supplement' for more details, but it does not explicitly state the dataset splits (e.g., train/validation/test percentages or counts) within the main text.
Hardware Specification No The paper does not specify any particular hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper does not provide specific software dependencies or version numbers for any libraries or frameworks used.
Experiment Setup No The paper states that 'more details on the datasets, training procedure, and hyperparameter settings can be found in Sections F and G of the supplement,' but these specific details are not provided in the main text.