DiffMD: A Geometric Diffusion Model for Molecular Dynamics Simulations

Authors: Fang Wu, Stan Z. Li

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

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct comprehensive experiments on multiple standard MD simulation datasets including MD17 and C7O2H10 isomers. Numerical results demonstrate that DIFFMD constantly outperforms state-of-the-art DLMD models by a large margin.
Researcher Affiliation Academia Fang Wu1,2, Stan Z. Li1* 1 AI Research and Innovation Laboratory, School of Engineering, Westlake University 2 Institute of AI Industry Research, Tsinghua University fw2359@columbia.edu, stan.zq.li@westlake.edu.cn
Pseudocode Yes Algorithm 1: Sampling Algorithm with Predictor-Corrector.
Open Source Code No The paper mentions dataset availability but does not provide concrete access to source code for the described methodology.
Open Datasets Yes MD17 (Chmiela et al. 2017) and C7O2H10 (Brockherde et al. 2017) datasets are available at http://quantum-machine.org/datasets/
Dataset Splits Yes We use the first 20K frame pairs as the training set and split the next 20K frame pairs equally into validation and test sets.
Hardware Specification No The paper does not specify the hardware (e.g., GPU/CPU models) used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup No The paper states 'The full experimental details are elaborated in Appendix.' and does not provide specific experimental setup details such as hyperparameter values or training configurations in the main text.