动态调整采样数量，让GNN训练更快更准
DAFOS: Dynamic Adaptive Fanout Optimization Sampler
- 根据节点重要性动态调整邻居采样数
- ogbn-arxiv上快3.57倍，准确率提升2.71个百分点
- 适合大规模图神经网络训练场景
图神经网络（GNN）在处理图结构数据时日益重要，但传统的均匀邻居采样和固定扇出设置常限制其可扩展性和效率。本文提出动态自适应扇出优化采样器（DAFOS），通过模型性能反馈动态调整扇出，并基于节点度对重要节点进行优先采样。随着训练进程推进，算法逐步增加扇出值，同时引入早期停止机制以在性能收益下降时终止训练。在ogbn-arxiv、Reddit和ogbn-products三个基准数据集上的实验表明，DAFOS显著提升了训练速度与精度：在ogbn-arxiv上实现3.57倍加速，准确率从68.5%提升至71.21%；在Reddit上实现12.6倍加速；在ogbn-products上准确率从73.78%提升至76.88%。结果验证了DAFOS在大规模GNN训练中的高效性与可扩展性。
原文摘要 · Abstract (English)
Graph Neural Networks (GNNs) are becoming an essential tool for learning from graph-structured data, however uniform neighbor sampling and static fanout settings frequently limit GNNs' scalability and efficiency. In this paper, we propose the Dynamic Adaptive Fanout Optimization Sampler (DAFOS), a novel approach that dynamically adjusts the fanout based on model performance and prioritizes important nodes during training. Our approach leverages node scoring based on node degree to focus computational resources on structurally important nodes, incrementing the fanout as the model training progresses. DAFOS also integrates an early stopping mechanism to halt training when performance gains diminish. Experiments conducted on three benchmark datasets, ogbnarxiv, Reddit, and ogbn-products, demonstrate that our approach significantly improves training speed and accuracy compared to a state-of-the-art approach. DAFOS achieves a 3.57x speedup on the ogbn-arxiv dataset and a 12.6x speedup on the Reddit dataset while improving the F1 score from 68.5% to 71.21% on ogbn-arxiv and from 73.78% to 76.88% on the ogbn-products dataset, respectively. These results highlight the potential of DAFOS as an efficient and scalable solution for large-scale GNN training.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。