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..
Unsupervised Federated Graph Learning
Authors: Lele Fu, Tianchi Liao, Sheng Huang, Bowen Deng, zhangchuanfu, Shirui Pan, Chuan Chen
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on eight graph datasets are conducted, the results demonstrate that the proposed Fed PAM is superior over classical and SOTA compared methods. |
| Researcher Affiliation | Academia | 1Sun Yat-sen University, Guangzhou, China 2Griffith University, Brisbane, Australia |
| Pseudocode | Yes | C The Algorithm Flow of the Proposed Fed PAM Algorithm 1 The main steps of Fed PAM ... Algorithm 4 The main steps for low-rank tensor optimization |
| Open Source Code | No | Question: Does the paper provide open access to the data and code, with sufficient instructions to faithfully reproduce the main experimental results, as described in supplemental material? Answer: [Yes] Justification: The algorithm flow, used datasets, and the parameter settings are introduced in details. |
| Open Datasets | Yes | Eight graph datasets are selected as benchmark datasets, including Cora, Cite Seer, Pub Med, Ogbn-Arxiv, Computers, Photo, Physics, Amazon-ratings. |
| Dataset Splits | Yes | Table 4: Descriptions of eight graph datasets. ... Train / Val / Test Category ... Cora 2,708 1,433 5,429 7 20% / 40% / 40% Citation Network |
| Hardware Specification | No | Question: For each experiment, does the paper provide sufficient information on the computer resources (type of compute workers, memory, time of execution) needed to reproduce the experiments? Answer: [Yes] Justification: We report the configuration of the server used in the implementation details. |
| Software Dependencies | No | A 2-layer graph convolutional network [62] is used as the backbone, and the latent embedding dimension is set as 128. ... We use Adam as the optimizer with the learning rate set to 0.001. |
| Experiment Setup | Yes | A 2-layer graph convolutional network [62] is used as the backbone, and the latent embedding dimension is set as 128. ... We use Adam as the optimizer with the learning rate set to 0.001. The numbers of communication round and local training rounds are set to 100 and 5, respectively. ... λ is tuned in {0.1, 5, 10, 50, 70, 100}, β is varied in {0.1, 0.5, 1}, the number of anchors is tuned in {30, 60, 100, 500, 1000}. |