A Graph Foundation Model for Large-Scale MIMO Detection
Abstract
Large-scale multiple-input multiple-output (MIMO) detection is fundamental to modern wireless networks but constrained by performance-complexity trade-offs. Existing detectors, whether classical or learning-based, often fall short in either scalability or generalizability across heterogeneous scenarios. To overcome these limitations, we introduce a wireless-native graph foundation model (GFM) tailored for large-scale MIMO detection. The proposed GFM employs a physics-informed hybrid architecture, integrating the local correlation extraction of message passing neural networks with the global attention of graph Transformers, encoding the physical interference patterns from the expectation propagation algorithm. Via extensive pre-training, this synergy enables the learning of a general-purpose detection mapping scalable across antenna dimensions and channel conditions. For rapid downstream deployment, parameter-efficient fine-tuning is leveraged to adapt the GFM to specific non-ideal system regimes with minimal overhead. To enhance inference efficiency, a mixture-of-experts mechanism is embedded at downstream deployment to dynamically activate only the necessary sub-modules. Evaluations show that the proposed GFM consistently outperforms classical detectors and advanced data-driven baselines in accuracy, configuration generality, and cross-scenario transferability across various challenging zero-shot and few-shot conditions.
Related
- Hybrid-Field Channel Tracking for Extremely Large-Scale MIMO Systems with Mobility
- Tri-Hybrid Beamforming Design for Large-Scale MIMO ISAC Systems
- Group Behavior-Guided Multi-Density Graph Clustering for Large-Scale Telemarketing Recommendation
- Resilient Cell-Free MIMO ISAC Under Uncertainties: An Impairment-Aware Graph Neural Network Framework
- Study of Iterative Detection and Decoding for {Mixed Near- and Far-Field} XL-MIMO Systems
BibTeX
@misc{zhou2026graph,
title = {{A Graph Foundation Model for Large-Scale MIMO Detection}},
author = {Xingyu Zhou and Le Liang and Hao Ye and Jing Zhang and Chao-Kai Wen and Xiao Li and Shi Jin and Wei Zhang},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.06361},
eprint = {2609.06361},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.06361},
}
RIS
TY - UNPB
TI - A Graph Foundation Model for Large-Scale MIMO Detection
AU - Xingyu Zhou
AU - Le Liang
AU - Hao Ye
AU - Jing Zhang
AU - Chao-Kai Wen
AU - Xiao Li
AU - Shi Jin
AU - Wei Zhang
PY - 2026
DA - 2026/09/06
PB - arXiv
DO - 10.48550/arxiv.2609.06361
UR - https://arxiv.org/abs/2609.06361
AB - Large-scale multiple-input multiple-output (MIMO) detection is fundamental to modern wireless networks but constrained by performance-complexity trade-offs. Existing detectors, whether classical or learning-based, often fall short in either scalability or generalizability across heterogeneous scenarios. To overcome these limitations, we introduce a wireless-native graph foundation model (GFM) tailored for large-scale MIMO detection. The proposed GFM employs a physics-informed hybrid architecture, integrating the local correlation extraction of message passing neural networks with the global attention of graph Transformers, encoding the physical interference patterns from the expectation propagation algorithm. Via extensive pre-training, this synergy enables the learning of a general-purpose detection mapping scalable across antenna dimensions and channel conditions. For rapid downstream deployment, parameter-efficient fine-tuning is leveraged to adapt the GFM to specific non-ideal system regimes with minimal overhead. To enhance inference efficiency, a mixture-of-experts mechanism is embedded at downstream deployment to dynamically activate only the necessary sub-modules. Evaluations show that the proposed GFM consistently outperforms classical detectors and advanced data-driven baselines in accuracy, configuration generality, and cross-scenario transferability across various challenging zero-shot and few-shot conditions.
ER -