TopGQ: Fast GNN Post-Training Quantization Leveraging Topology Information
2026-08-31 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors address the problem of slow and inefficient quantization in graph neural networks (GNNs), which makes them hard to use in practice. They introduce TopGQ, a method that speeds up quantization by reducing unnecessary work. Their technique includes dual-axis scale absorption, which simplifies how activations are quantized, and TopPIN, which groups similar nodes to make quantization more efficient. Tests show that TopGQ cuts quantization time drastically while keeping accuracy intact.
Graph Neural NetworksQuantizationPost-training QuantizationActivation QuantizationAdjacency MatrixNode GroupingGraph TopologyComputational EfficiencyMachine Learning Optimization
Authors
Dain Kwon, Kanghyun Choi, Hyeyoon Lee, Sunjong Park, Seoyong Lee, Sukjin Kim, Jinho Lee
Abstract
Existing GNN quantization methods suffer from considerable quantization overhead, which severely limits their practical usage in real-world scenarios. To this end, we present TopGQ, an accurate post-training GNN quantization framework, alleviating redundant quantization overhead. We propose dual-axis scale absorption, which enables activation quantization along both the outer and inner dimensions by merging one into the adjacency matrix. On top of that, we introduce TopPIN, a proxy for nodes' local structure, and use it to group nodes with similar topology during quantization. Experimental results show that TopGQ reduces quantization time by an order of magnitude while preserving accuracy.