Lune

ICLR2023Top-tier venue

A2Q\rm A^2Q: Aggregation-Aware Quantization for Graph Neural Networks

Zeyu Zhu, Fanrong Li, Zitao Mo, Qinghao Hu, Gang Li, Zejian Liu, Xiaoyao Liang, Jian Cheng

Abstract

As graph data size increases, the vast latency and memory consumption during inference pose a significant challenge to the real-world deployment of Graph Neural Networks (GNNs). While quantization is a powerful approach to reducing GNNs complexity, most previous works on GNNs quantization fail to exploit the unique characteristics of GNNs, suffering from severe accuracy degradation. Through an in-depth analysis of the topology of GNNs, we observe that the topology of the graph leads to significant differences between nodes, and most of the nodes in a graph appear to have a small aggregation value. Motivated by this, in this paper, we propose the Aggregation-Aware mixed-precision Quantization (A 2 Q) for GNNs, where an appropriate bitwidth is automatically learned and assigned to each node in the graph. To mitigate the vanishing gradient problem caused by sparse connections between nodes, we propose a Local Gradient method to serve the quantization error of the node features as the supervision during training. We also develop a Nearest Neighbor Strategy to deal with the generalization on unseen graphs. Extensive experiments on eight public node-level and graph-level datasets demonstrate the generality and robustness of our proposed method. Compared to the FP32 models, our method can achieve up to a 18.6x (i.e., 1.70bit) compression ratio with negligible accuracy degradation. Morever, compared to the state-of-theart quantization method, our method can achieve up to 11.4% and 9.5% accuracy improvements on the node-level and graph-level tasks, respectively, and up to 2x speedup on a dedicated hardware accelerator.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

1 top-tier paper has cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • Open Graph Benchmark: Datasets for Machine Learning on Graphs cites GraphSAINT: Graph Sampling Based Inductive Learning Method
  • HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks cites HAWQ: Hessian AWare Quantization of Neural Networks With Mixed-Precision
  • Towards Efficient and Accurate Winograd Convolution via Full Quantization cites Learned Step Size quantization
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers1

  1. Towards Efficient and Accurate Winograd Convolution via Full QuantizationTianqi Chen, Weixiang Xu, Weihan Chen, Peisong Wang et al.NeurIPS 2023 · 13 citations

Builds on14

  1. Open Graph Benchmark: Datasets for Machine Learning on GraphsWeihua Hu, Matthias Fey, Marinka Zitnik, Yuxiao Dong et al.NeurIPS 2020 · 3,935 citations
  2. GraphSAINT: Graph Sampling Based Inductive Learning MethodHanqing Zeng, Hongkuan Zhou, Ajitesh Srivastava, Rajgopal Kannan et al.ICLR 2020 · 1,155 citations
  3. Learned Step Size quantizationSteven K. Esser, Jeffrey L. McKinstry, Deepika Bablani, Rathinakumar Appuswamy et al.ICLR 2020 · 1,037 citations
  4. HAWQ: Hessian AWare Quantization of Neural Networks With Mixed-PrecisionZhen Dong, Zhewei Yao, Amir Gholami, Michael W. Mahoney et al.ICCV 2019 · 645 citations
  5. HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural NetworksZhen Dong, Zhewei Yao, Daiyaan Arfeen, Amir Gholami et al.NeurIPS 2020 · 434 citations

Related papers

  1. Efficient Mixed Precision Quantization in Graph Neural NetworksSamir Moustafa, Nils M. Kriege, Wilfried N. GanstererICDE 2025 · 2 citations
  2. MEGA: A Memory-Efficient GNN Accelerator Exploiting Degree-Aware Mixed-Precision QuantizationZeyu Zhu, Fanrong Li, Gang Li, Zejian Liu et al.HPCA 2024 · 28 citations
  3. NodeBits: A Plug-and-Play Framework for Accelerating Graph Inference by Post-Hoc Binary QuantizationQihao Cheng, Tianhao Wu, Da Yan, Haoran TangKDD 2026
  4. Degree-Quant: Quantization-Aware Training for Graph Neural NetworksShyam Anil Tailor, Javier Fernández-Marqués, Nicholas Donald LaneICLR 2021 · 180 citations
  5. FLAG: An FPGA-Based System for Low-Latency GNN Inference Service Using Vector QuantizationYunki Han, Taehwan Kim, Jiwan Kim, Seohye Ha et al.DAC 2025 · 1 citation
Dusk over the sea between two cliffs drawn in fine vertical lines