A Prior-based Discrete Diffusion Model for Social Graph Generation

Authors: Shu Yin, Dongpeng Hou, Lianwei Wu, Xianghua Li, Chao Gao

IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Comprehensive experiments on real-world social network datasets demonstrate PDDM s superiority in terms of the MMD metric and downstream tasks.
Researcher Affiliation Academia Shu Yin1,2 , Dongpeng Hou3,2 , Lianwei Wu1 , Xianghua Li2 and Chao Gao2 1School of Computer Science, Northwestern Polytechnical University 2School of Artificial Intelligence, OPtics and Electro Nics (i OPEN), Northwestern Polytechnical University 3School of Mechanical Engineering, Northwestern Polytechnical University EMAIL
Pseudocode Yes Algorithm 1 Training for PDDM Input: A total of K propagation graphs Gk = (Vk, Ek, Fk). Output: Optimized parameters θ. 1: repeat ... Algorithm 2 Graph generation process of PDDM Input: The optimized denoising module θ; A candidate user set (V, F). Output: The generative graph G. 1: Determine the number of the diffusion step T = |V |
Open Source Code Yes The code is available at https://github.com/cgao-comp/PDDM.
Open Datasets Yes We use real-world propagation graphs on Weibo and Twitter platforms for graph generation, namely Weibo [Ma et al., 2017], Twitter15, and Twitter16 [Liu et al., 2015; Ma et al., 2016].
Dataset Splits Yes Following the SOTA settings, we use 80% of the graphs as training set and the rest 20% as test sets [Kong et al., 2023] for each dataset.
Hardware Specification No The paper does not provide specific hardware details such as exact GPU/CPU models, processor types, or memory amounts used for running its experiments.
Software Dependencies No The paper mentions using GCNs and GATs as denoising modules but does not provide specific software library names or version numbers (e.g., PyTorch, TensorFlow, or Python versions) needed to replicate the experiment.
Experiment Setup No The paper specifies dataset splits and evaluation metrics in Section 4.2 'Experimental Setting', but it does not provide concrete hyperparameter values such as learning rates, batch sizes, number of epochs, or specific optimizer settings for training the models.