BLISS: Bandit Layer Importance Sampling Strategy for Efficient Training of Graph Neural Networks
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918264463949824 |
|---|---|
| author | Alsaqa, Omar Hoang, Linh Thi Balin, Muhammed Fatih |
| author_facet | Alsaqa, Omar Hoang, Linh Thi Balin, Muhammed Fatih |
| contents | Graph Neural Networks (GNNs) are powerful tools for learning from graph-structured data, but their application to large graphs is hindered by computational costs. The need to process every neighbor for each node creates memory and computational bottlenecks. To address this, we introduce BLISS, a Bandit Layer Importance Sampling Strategy. It uses multi-armed bandits to dynamically select the most informative nodes at each layer, balancing exploration and exploitation to ensure comprehensive graph coverage. Unlike existing static sampling methods, BLISS adapts to evolving node importance, leading to more informed node selection and improved performance. It demonstrates versatility by integrating with both Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs), adapting its selection policy to their specific aggregation mechanisms. Experiments show that BLISS maintains or exceeds the accuracy of full-batch training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_22388 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BLISS: Bandit Layer Importance Sampling Strategy for Efficient Training of Graph Neural Networks Alsaqa, Omar Hoang, Linh Thi Balin, Muhammed Fatih Machine Learning Artificial Intelligence Social and Information Networks Optimization and Control 68T05, 05C85, 62L05, 68T07 I.2.6; G.2.2; F.2.2 Graph Neural Networks (GNNs) are powerful tools for learning from graph-structured data, but their application to large graphs is hindered by computational costs. The need to process every neighbor for each node creates memory and computational bottlenecks. To address this, we introduce BLISS, a Bandit Layer Importance Sampling Strategy. It uses multi-armed bandits to dynamically select the most informative nodes at each layer, balancing exploration and exploitation to ensure comprehensive graph coverage. Unlike existing static sampling methods, BLISS adapts to evolving node importance, leading to more informed node selection and improved performance. It demonstrates versatility by integrating with both Graph Convolutional Networks (GCNs) and Graph Attention Networks (GATs), adapting its selection policy to their specific aggregation mechanisms. Experiments show that BLISS maintains or exceeds the accuracy of full-batch training. |
| title | BLISS: Bandit Layer Importance Sampling Strategy for Efficient Training of Graph Neural Networks |
| topic | Machine Learning Artificial Intelligence Social and Information Networks Optimization and Control 68T05, 05C85, 62L05, 68T07 I.2.6; G.2.2; F.2.2 |
| url | https://arxiv.org/abs/2512.22388 |