DBScholar

Back to papers

Faster Convergence in Mini-batch Graph Neural Networks Training with Pseudo Full Neighborhood Compensation

Summary: Identifies two gradient-estimation error sources in mini-batch neighbor-sampled GNNs—missing contributions from unsampled target nodes and inaccuracies in sampled-node messages—and shows prior work largely ignores the former. Proposes PFNC: a history-based, partial-cache, sampler-agnostic compensation scheme that corrects both errors, provably better approximates full-batch gradients, and empirically speeds convergence and improves generalization. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID
h74e3dc72c08e9157
Venue
VLDB
Year
2025
Pagerank
4.9793485e-05
Overall Rank
11,361 | 23.62%
DOI
10.14778/3749646.3749695

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

BibTeX Citation

@article{zhou_vldb25,
        title = {{Faster Convergence in Mini-batch Graph Neural Networks Training with Pseudo Full Neighborhood Compensation}},
        author = {Zhou, Qiqi and Shen, Yanyan and Chen, Lei},
        journal = {PVLDB},
        series = {{VLDB} '25},
        volume = {18},
        number = {11},
        pages = {4309--4322},
        doi = {10.14778/3749646.3749695},
        url = {https://doi.org/10.14778/3749646.3749695},
        year = {2025}
}

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank Citing Paper Year Venue Pagerank
10,706 Efficient GNN Training on Giant Graphs with Collective Batching and Scheduling 2026 VLDB 4.9793485e-05
Previous Page 1 / 1 Next

Outgoing Citations (Sorted by Pagerank)

Showing 17 of 17 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank Cited Paper Year Venue Pagerank
211 AliGraph: A Comprehensive Graph Neural Network Platform 2019 VLDB 0.00024816965
1,134 SANCUS: Staleness-Aware Communication-Avoiding Full-Graph Decentralized Training in Large-Scale Graph Neural Networks 2022 VLDB 0.00011893521
2,506 DUCATI: A Dual-Cache Training System for Graph Neural Networks on Giant Graphs with the GPU 2023 SIGMOD 8.3774747e-05
2,596 Scalable and Efficient Full-Graph GNN Training for Large Graphs 2023 SIGMOD 8.2419456e-05
4,502 NeutronStream: A Dynamic GNN Training Framework with Sliding Window for Graph Streams 2024 VLDB 6.5740722e-05
4,764 ETC: Efficient Training of Temporal Graph Neural Networks over Large-scale Dynamic Graphs 2024 VLDB 6.4276208e-05
4,825 DAHA: Accelerating GNN Training with Data and Hardware Aware Execution Planning 2024 VLDB 6.3923554e-05
4,853 NeutronOrch: Rethinking Sample-based GNN Training under CPU-GPU Heterogeneous Environments 2024 VLDB 6.3805655e-05
4,898 Comprehensive Evaluation of GNN Training Systems: A Data Management Perspective 2024 VLDB 6.365824e-05
4,993 DGC: Training Dynamic Graphs with Spatio-Temporal Non-Uniformity using Graph Partitioning by Chunks 2023 SIGMOD 6.32357e-05
5,310 DynaHB: A Communication-Avoiding Asynchronous Distributed Framework with Hybrid Batches for Dynamic GNN Training 2024 VLDB 6.1857447e-05
5,339 FreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training 2024 VLDB 6.1731674e-05
5,899 Lotan: Bridging the Gap between GNNs and Scalable Graph Analytics Engines 2023 VLDB 5.9545431e-05
6,780 Eliminating Data Processing Bottlenecks in GNN Training over Large Graphs via Two-level Feature Compression 2024 VLDB 5.6838458e-05
6,826 SIMPLE: Efficient Temporal Graph Neural Network Training at Scale with Dynamic Data Placement 2024 SIGMOD 5.6700316e-05
6,924 HongTu: Scalable Full-Graph GNN Training on Multiple GPUs 2023 SIGMOD 5.6426299e-05
7,506 ADGNN: Towards Scalable GNN Training with Aggregation-Difference Aware Sampling 2023 SIGMOD 5.5070363e-05
Previous Page 1 / 1 Next

Semantically Similar Papers

# Overall Rank Paper Year Venue
1 1,772 ByteGNN: Efficient Graph Neural Network Training at Large Scale 2022 VLDB
2 7,506 ADGNN: Towards Scalable GNN Training with Aggregation-Difference Aware Sampling 2023 SIGMOD
3 6,780 Eliminating Data Processing Bottlenecks in GNN Training over Large Graphs via Two-level Feature Compression 2024 VLDB
4 8,630 Historical Embedding-Guided Efficient Large-Scale Federated Graph Learning 2024 SIGMOD
5 6,324 EARLY: Efficient and Reliable Graph Neural Network for Dynamic Graphs 2023 SIGMOD
6 5,339 FreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training 2024 VLDB
7 4,898 Comprehensive Evaluation of GNN Training Systems: A Data Management Perspective 2024 VLDB
8 2,596 Scalable and Efficient Full-Graph GNN Training for Large Graphs 2023 SIGMOD
9 6,669 Efficient Training of Graph Neural Networks on Large Graphs 2024 VLDB
10 10,257 Graph Neural Network Training Systems: A Performance Comparison of Full-Graph and Mini-Batch 2025 VLDB