Heta: Distributed Training of Heterogeneous Graph Neural Networks
Summary: Distributed HGNN training is communication-bound due to per-type feature dims and featureless nodes. Heta: relation-first aggregation, schema-aware meta-partitioning, and type-aware GPU cache cut cross-machine communication and yield ~5.3× speedups on large HetGs. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Yuchen Zhong (University of Hong Kong)
- 2. Junwei Su (University of Hong Kong)
- 3. Chuan Wu (University of Hong Kong)
- 4. Minjie Wang (AWS Shanghai AI Lab)
BibTeX Citation
@article{zhong_vldb25,
title = {{Heta: Distributed Training of Heterogeneous Graph Neural Networks}},
author = {Zhong, Yuchen and Su, Junwei and Wu, Chuan and Wang, Minjie},
journal = {PVLDB},
series = {{VLDB} '25},
volume = {18},
number = {9},
pages = {2790--2803},
doi = {10.14778/3746405.3746408},
url = {https://doi.org/10.14778/3746405.3746408},
year = {2025}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 8 of 8 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 3 | Pregel: A System for Large-Scale Graph Processing | 2010 | SIGMOD | 0.0012250108 |
| 223 | AliGraph: A Comprehensive Graph Neural Network Platform | 2019 | VLDB | 0.00024182473 |
| 521 | PyTorch Distributed: Experiences on Accelerating Data Parallel Training | 2020 | VLDB | 0.0001713368 |
| 1,132 | SANCUS: Staleness-Aware Communication-Avoiding Full-Graph Decentralized Training in Large-Scale Graph Neural Networks | 2022 | VLDB | 0.00012041292 |
| 1,863 | ByteGNN: Efficient Graph Neural Network Training at Large Scale | 2022 | VLDB | 9.5950349e-05 |
| 5,243 | FreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training | 2024 | VLDB | 6.3018825e-05 |
| 5,316 | Comprehensive Evaluation of GNN Training Systems: A Data Management Perspective | 2024 | VLDB | 6.2687017e-05 |
| 6,803 | OUTRE: An OUT-of-core De-REdundancy GNN Training Framework for Massive Graphs within A Single Machine | 2024 | VLDB | 5.7684339e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 1,863 | ByteGNN: Efficient Graph Neural Network Training at Large Scale | 2022 | VLDB |
| 2 | 2,162 | Heterogeneity-aware Distributed Parameter Servers | 2017 | SIGMOD |
| 3 | 6,806 | HongTu: Scalable Full-Graph GNN Training on Multiple GPUs | 2023 | SIGMOD |
| 4 | 5,760 | DynaHB: A Communication-Avoiding Asynchronous Distributed Framework with Hybrid Batches for Dynamic GNN Training | 2024 | VLDB |
| 5 | 4,956 | Heterogeneity-Aware Distributed Machine Learning Training via Partial Reduce | 2021 | SIGMOD |
| 6 | 9,729 | Scalable Graph Convolutional Network Training on Distributed-Memory Systems | 2023 | VLDB |
| 7 | 2,640 | Scalable and Efficient Full-Graph GNN Training for Large Graphs | 2023 | SIGMOD |
| 8 | 2,485 | HET: Scaling out Huge Embedding Model Training via Cache-enabled Distributed Framework | 2022 | VLDB |
| 9 | 4,912 | HET-GMP: A Graph-based System Approach to Scaling Large Embedding Model Training | 2022 | SIGMOD |
| 10 | 10,323 | NeutronHeter: Optimizing Distributed Graph Neural Network Training for Heterogeneous Clusters | 2026 | SIGMOD |