Graph Learning in 4D: A Quaternion-Valued Laplacian to Enhance Spectral GCNs
Authors: Stefano Fiorini, Stefano Coniglio, Michele Ciavotta, Enza Messina
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical Experiments We compare Quater GCN with state-of-the-art GCNs across four tasks: node classification (NC), three-class edge prediction (3CEP), four-class edge prediction (4CEP), and five-class edge prediction (5CEP). Throughout this section, the tables report the best results in boldface and the second-best are underlined. The datasets and the code we used are publicly available at https://github.com/Stefa1994/Quater GCN. We experiment on the six widely-used real-world directed graphs Bitcoin-OTC, Bitcoin Alpha, Wiki Rfa, Telegram, Slashdot, and Epinions |
| Researcher Affiliation | Academia | 1 Italian Institute of Technology, Genoa, Italy 2 University of Bergamo, Bergamo, Italy 3 University of Milano-Bicocca, Milan, Italy |
| Pseudocode | No | The paper describes the mathematical formulation and structure of Quater GCN, but it does not include explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | The datasets and the code we used are publicly available at https://github.com/Stefa1994/Quater GCN. |
| Open Datasets | Yes | We experiment on the six widely-used real-world directed graphs Bitcoin-OTC, Bitcoin Alpha, Wiki Rfa, Telegram, Slashdot, and Epinions (see Kumar et al. (2016); Bovet and Grindrod (2020); West et al. (2014); Leskovec, Huttenlocher, and Kleinberg (2010)). |
| Dataset Splits | Yes | The experiments are run with 10-fold cross-validation with a 60%/20%/20% split for training, validation, and testing. |
| Hardware Specification | No | The paper mentions calculations on a 'GPU' in a general sense when discussing complexity, but it does not provide specific details on the hardware used for the experiments, such as exact GPU or CPU models, memory, or cloud infrastructure. |
| Software Dependencies | No | The paper mentions 'Py Torch Geometric Signed Directed' as a related software package but does not list specific version numbers for software dependencies like Python, PyTorch, or other libraries used in their implementation. |
| Experiment Setup | No | The paper describes the general architecture and some processing steps of Quater GCN, but it does not provide specific experimental setup details such as hyperparameter values (e.g., learning rate, batch size, number of epochs), optimizer settings, or detailed training configurations. |