D3-GNN 实现了实时图神经网络的高效流式更新，支持高吞吐与低延迟。
D3-GNN: Dynamic Distributed Dataflow for Streaming Graph Neural Networks
- 采用流水线式计算图与窗口化前向传播处理动态图更新
- 相比 DGL，吞吐提升约 76 倍，运行时间减少 10 倍，消息量降低 15 倍
- 适合大规模实时图推理场景，尤其在数据倾斜和邻居爆炸时表现优异
流式图神经网络（GNN）在持续捕捉图的动态状态方面面临算法挑战，同时在推理与训练中需优化延迟、内存与吞吐。我们提出 D3-GNN，首个面向在线查询场景的分布式、混合并行流式 GNN 系统。该系统解决数据管理、算法与系统难题，实现图状态的连续捕获与节点表示的容错更新，具备最优延迟、负载均衡与吞吐。D3-GNN 采用流式 GNN 聚合器与展开的分布式计算图架构，应对级联图更新。为缓解数据偏斜与邻居爆炸问题，引入层间与层内窗口化前向传播机制。在大规模图流上的实验表明，D3-GNN 具有高效率与可扩展性：相比 DGL，流式任务下吞吐提升约 76 倍；窗口化改进进一步使运行时间减少约 10 倍，消息量最多降低 15 倍，且在更高并行度下仍保持高效。
原文摘要 · Abstract (English)
Graph Neural Network (GNN) models on streaming graphs entail algorithmic challenges to continuously capture its dynamic state, as well as systems challenges to optimize latency, memory, and throughput during both inference and training. We present D3-GNN, the first distributed, hybrid-parallel, streaming GNN system designed to handle real-time graph updates under online query setting. Our system addresses data management, algorithmic, and systems challenges, enabling continuous capturing of the dynamic state of the graph and updating node representations with fault-tolerance and optimal latency, load-balance, and throughput. D3-GNN utilizes streaming GNN aggregators and an unrolled, distributed computation graph architecture to handle cascading graph updates. To counteract data skew and neighborhood explosion issues, we introduce inter-layer and intra-layer windowed forward pass solutions. Experiments on large-scale graph streams demonstrate that D3-GNN achieves high efficiency and scalability. Compared to DGL, D3-GNN achieves a significant throughput improvement of about 76x for streaming workloads. The windowed enhancement further reduces running times by around 10x and message volumes by up to 15x at higher parallelism.
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。