A Graph Dynamics Prior for Relational Inference

Authors: Liming Pan, Cheng Shi, Ivan Dokmanic

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments show that GDP reconstructs graphs far more accurately than earlier methods, with remarkable robustness to under-sampling. Experiments show that GDP achieves significantly higher inference accuracy than any of the earlier approaches. We consider several representative graph dynamical systems, both continuous and discrete, to validate the proposed algorithm. The results for NRI and GDP are averaged over 10 independent runs. Table 1 summarizes the average AUC with standard deviations.
Researcher Affiliation Academia 1School of Cyber Science and Technology, University of Science and Technology of China, Hefei, China 230026, 2School of Computer and Electronic Information, Nanjing Normal University, China 210023, 3 Departement Mathematik und Informatik, Universit at Basel, Basel, Switzerland 4051
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks. The model architecture is described textually and with a diagram.
Open Source Code Yes Reproducible code is available at https://github.com/Da Da Cheng/GDP.
Open Datasets Yes We consider several representative graph dynamical systems, both continuous and discrete, to validate the proposed algorithm. The continuous-time systems include (i) the Michaelis Menten kinetics (Karlebach and Shamir 2008)... (v) the Kuramoto model (Kuramoto 1975)... Moreover, we considered a publicly available f MRI dataset (viii) Netsim (Smith et al. 2011), comprising realistic simulated data.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) needed to reproduce the data partitioning for training, validation, and testing. It mentions data volume and averaging over runs, but not explicit splits.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes The hyperparameters for GDP are summarized in Appendix B.3.