Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Multi-Scale Subgraph Contrastive Learning
Authors: Yanbei Liu, Yu Zhao, Xiao Wang, Lei Geng, Zhitao Xiao
IJCAI 2023 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments and parametric analysis on eight graph classification real-world datasets well demonstrate the effectiveness of the proposed method. |
| Researcher Affiliation | Academia | 1School of Life Sciences, Tiangong University 2School of Electronics and Information Engineering, Tiangong University 3School of Software, Beihang University |
| Pseudocode | Yes | Algorithm 1 The training process of the MSSGCL |
| Open Source Code | No | The paper does not provide any links or explicit statements about the availability of open-source code for the described methodology. |
| Open Datasets | Yes | We adopt the TUDataset benchmark [Morris et al., 2020], which contains different types of graphs, i.e., molecules and social networks |
| Dataset Splits | Yes | We use 10-fold cross validation accuracy to report classification performance. Experiments are repeated 5 times. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware (e.g., GPU models, CPU types) used for running its experiments. |
| Software Dependencies | No | The paper mentions using GIN as the encoder and Adam optimizer but does not specify version numbers for these or other software dependencies. |
| Experiment Setup | Yes | In our framework, we set the global view size to be 80% of the whole and the local view size to be 20% of the whole for molecular graphs, and 90% of the global view size and 10% of the local view size for social networks. The measurement function between local views is composed of a 5-layer MLPs with batch normalization and RELU activation functions. Its output is fed into a Sigmoid function, which outputs a scalar to indicate the similarity between two local views. [...] we adopt GIN as the encoder, and a sum pooling is used as the readout function. [...] where λ1 and λ2 are hyper-parameters to balance different loss terms. |