EUVIMEDEUVIMED entdecken
Crossref

GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting

Yuqi Yuan, Xiong Luo, Qiaojuan Peng, Wenbing Zhao

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (33) · 28041-28049

DOI: 10.1609/aaai.v40i33.40029 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Yuqi Yuan und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large Language Models (LLMs) have recently emerged as a leading approach for multivariate time series forecasting. However, their effectiveness is hampered by a fundamental architectural mismatch: the permutation-invariant self-attention of Transformers lacks inductive biases for the strict temporal order and complex cross-variable dependencies inherent in time series. Existing methods often sidestep this issue with input-level alignment techniques rather than endowing the model itself with structural awareness. To address this gap, we introduce GraFT (Graph-infused Forecasting Transformer), a framework that systematically embeds relational priors into a pre-trained backbone by constructing a heterogeneous patch relation graph, which represents both universal temporal principles with static edges and instance-specific patterns with dynamic adaptive edges. To process this multi-relational structure, a relational graph convolutional network generates structure-aware representations, which are infused into the patch embeddings to provide explicit structural guidance to the Transformer's attention mechanism. Extensive experiments show that GraFT achieves state-of-the-art performance on long-term forecasting and zero-shot learning, outperforming leading LLM-based methods on eight standard benchmarks with an average Mean Squared Error (MSE) reduction of 14.4%.

Zitieren

APA

Yuan, Y., Luo, X., Peng, Q., & Zhao, W. (2026). GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 40(33), 28041-28049. https://doi.org/10.1609/aaai.v40i33.40029

Vancouver

Yuan Y, Luo X, Peng Q, Zhao W. GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(33):28041-28049. doi:10.1609/aaai.v40i33.40029

BibTeX
@article{Yuan2026,
  author = {Yuan, Yuqi and Luo, Xiong and Peng, Qiaojuan and Zhao, Wenbing},
  title = {GraFT: Infusing Pre-trained Transformers with Relational Structure for Time Series Forecasting},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {33},
  pages = {28041--28049},
  doi = {10.1609/aaai.v40i33.40029}
}
Kostenlos starten →