Adaptable graph region for optimizing performance in dynamic system long-term forecasting via time-aware expert

Xuan Peng, Zefeng Liu, Peng Zhang, Yufei Chen, Zhanjun Shao, Han Zhao, Xiaonan Xie, Lizhong Jiang, Zhuo Huang, Zhouzhou Pan, Jianwei Yan, Binbin Yin and Ping Xiang ()
Additional contact information
Xuan Peng: Central South University, School of Civil Engineering
Zefeng Liu: Central South University, School of Civil Engineering
Peng Zhang: Central South University, School of Civil Engineering
Yufei Chen: Central South University, School of Civil Engineering
Zhanjun Shao: Central South University, School of Civil Engineering
Han Zhao: Central South University, School of Civil Engineering
Xiaonan Xie: Central South University, School of Civil Engineering
Lizhong Jiang: Central South University, School of Civil Engineering
Zhuo Huang: Changsha University of Science and Technology, School of Civil Engineering
Zhouzhou Pan: University of Oxford, Department of Engineering Science
Jianwei Yan: East China Jiaotong University, School of Civil Engineering and Architecture
Binbin Yin: The Hong Kong Polytechnic University, Department of Civil and Environmental Engineering
Ping Xiang: Central South University, School of Civil Engineering

Nature Communications, 2025, vol. 16, issue 1, 1-16

Abstract: Abstract Real-time and accurate prediction of the long-term behavior of dynamic systems is crucial for identifying risks during unexpected events, while computational efficiency is significantly influenced by the scale of the dynamic system. However, existing neural network models mainly focus on optimizing network structures to improve accuracy, neglecting computational efficiency. To address this issue, we propose regional graph representation, which reduces the scale of the graph structure by merging nodes into region, extracting topological information through graph convolution or lightweight convolution modules, and restoring the regions via fine-grained reconstruction. Notably, this method is adaptable to all graph-based models. Meanwhile, we introduce a sparse time-aware expert module, which selects experts for processing different scale information through a dynamic sparse selection mechanism, enabling multi-scale modeling of temporal information. The architecture we achieve an optimal balance between speed and prediction accuracy, providing a practical solution for real-time forecasting.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.nature.com/articles/s41467-025-64984-w Abstract (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:nat:natcom:v:16:y:2025:i:1:d:10.1038_s41467-025-64984-w

Ordering information: This journal article can be ordered from
https://www.nature.com/ncomms/

DOI: 10.1038/s41467-025-64984-w

Access Statistics for this article

Nature Communications is currently edited by Nathalie Le Bot, Enda Bergin and Fiona Gillespie

More articles in Nature Communications from Nature
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().