Guided Graph Compression for Quantum Graph Neural Networks

2025-06-11

AI Summary

Quantum Graph Neural Networks (QGNNs) face fundamental bottlenecks on real-world large-scale graph data (e.g., Jet Tagging) due to quantum hardware constraints, limiting their ability to encode high-dimensional node features and dense topological structures. Method: We propose Guided Graph Compression (GGC), the first framework enabling joint, task-driven compression of both node count and feature dimensionality. GGC employs an end-to-end differentiable graph autoencoder, incorporates a classification-aware loss, and integrates a built-in QGNN ansatz validation interface to facilitate quantum-classical hybrid modeling. Contribution/Results: Unlike prior works relying on simplified or synthetic benchmarks, GGC achieves significant downstream classification performance gains on real physics data. It enables, for the first time, empirical evaluation of novel QGNN architectures on graphs with ~1,000 nodes—surpassing the limitations of manual preprocessing and unguided compression methods.

Technology Category

• Machine Learning: Quantum Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Graph Neural Networks (GNNs) are effective for processing graph-structured data but face challenges with large graphs due to high memory requirements and inefficient sparse matrix operations on GPUs. Quantum Computing (QC) offers a promising avenue to address these issues and inspires new algorithmic approaches. In particular, Quantum Graph Neural Networks (QGNNs) have been explored in recent literature. However, current quantum hardware limits the dimension of the data that can be effectively encoded. Existing approaches either simplify datasets manually or use artificial graph datasets. This work introduces the Guided Graph Compression (GGC) framework, which uses a graph autoencoder to reduce both the number of nodes and the dimensionality of node features. The compression is guided to enhance the performance of a downstream classification task, which can be applied either with a quantum or a classical classifier. The framework is evaluated on the Jet Tagging task, a classification problem of fundamental importance in high energy physics that involves distinguishing particle jets initiated by quarks from those by gluons. The GGC is compared against using the autoencoder as a standalone preprocessing step and against a baseline classical GNN classifier. Our numerical results demonstrate that GGC outperforms both alternatives, while also facilitating the testing of novel QGNN ansatzes on realistic datasets.

Problem

• Reducing large graph memory usage for quantum GNNs
• Overcoming quantum hardware data dimension limitations
• Enhancing classification performance in particle jet tagging

Innovation

• Guided Graph Compression reduces nodes and features
• Graph autoencoder enhances downstream classification performance
• Framework tested on Jet Tagging in high energy physics