aiqbench analysis
TopGQ addresses critical efficiency bottlenecks in Graph Neural Networks by integrating topology-aware quantization, significantly reducing post-training overhead without sacrificing accuracy. This advancement is particularly relevant for deploying GNNs on resource-constrained devices where inference speed and memory usage are paramount. By leveraging node local structure to group similar elements during quantization, the method offers a practical pathway toward scalable graph AI applications.
TopGQ: Fast GNN Post-Training Quantization Leveraging Topology Information
arXiv:2608.30394v1 Announce Type: cross 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.
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.
7 pages, 4 figures. Accepted at the 63rd ACM/IEEE Design Automation Conference (DAC 2026)
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2608.30394 [cs.LG]
(or arXiv:2608.30394v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.30394 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 208 words
