EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A spiking temporal gated framework with gated adaptive positional encoding for multivariate time-series forecasting

Fengge Wang, Mengzhao Zhang, Zhaoyuan Wang, Yongping Dan, Zhanyang Zhang

PeerJ Computer Science · 13.05.2026 · 12 · e3874

DOI: 10.7717/peerj-cs.3874 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PeerJ Computer Science, online erschienen am 13.05.2026. Verfasst von Fengge Wang und 4 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Henan Provincial Science and Technology Research Project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Spiking Neural Networks (SNNs) exhibit superior energy efficiency and biological plausibility due to their event-driven, sparse computations in temporal tasks. However, conventional convolution-based spiking models struggle to capture long-range dependencies and cross-channel interactions, thereby limiting their ability to represent complex temporal patterns. To address this limitation, this article introduces a parallel spiking temporal modeling framework that synergizes heterogeneous convolutional and gated recurrent architectures (Spiking Temporal Gated, STG) with adaptive positional encoding. Firstly, we design a parallel module consisting of two branches: spiking Heterogeneous Temporal Convolution Network (SHTCN), which fuses 1D causal convolutions with 2D cross-channel convolutions to extract short-term local features; spike-gated recurrent unit (GRU), a spiking-driven GRU that captures long-term global dependencies. Secondly, we propose Gated Adaptive Positional Encoding (GAPE), which employs learnable gating weights to dynamically modulate multi-frequency positional encodings, thereby enhancing the model’s sensitivity to temporal order and periodic structures. Extensive experiments on multiple multivariate time-series forecasting benchmarks demonstrate that STG-GAPE yields notable improvements in predictive accuracy and robustness compared with existing SNN-based approaches, while maintaining low computational complexity and a compact parameter footprint. These results validate the effectiveness of combining parallel heterogeneous modeling with adaptive positional encoding in SNNs.

Förderung

  • Henan Provincial Science and Technology Research Project

Zitieren

APA

Wang, F., Zhang, M., Wang, Z., Dan, Y., & Zhang, Z. (2026). A spiking temporal gated framework with gated adaptive positional encoding for multivariate time-series forecasting. PeerJ Computer Science, 12, e3874. https://doi.org/10.7717/peerj-cs.3874

Vancouver

Wang F, Zhang M, Wang Z, Dan Y, Zhang Z. A spiking temporal gated framework with gated adaptive positional encoding for multivariate time-series forecasting. PeerJ Computer Science. 2026;12:e3874. doi:10.7717/peerj-cs.3874

BibTeX
@article{Wang2026,
  author = {Wang, Fengge and Zhang, Mengzhao and Wang, Zhaoyuan and Dan, Yongping and Zhang, Zhanyang},
  title = {A spiking temporal gated framework with gated adaptive positional encoding for multivariate time-series forecasting},
  journal = {PeerJ Computer Science},
  year = {2026},
  volume = {12},
  pages = {e3874},
  doi = {10.7717/peerj-cs.3874}
}
Kostenlos starten →