Mixture of TSMixer Experts for Time Series Forecasting
Biomimetics · 15.06.2026 · 11 (6) · 426
DOI: 10.3390/biomimetics11060426 ↗Kurz gefasst
Fachartikel in der Zeitschrift Biomimetics, online erschienen am 15.06.2026. Verfasst von Jaemoo Hong und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, Chungbuk National University, Cheongju 28644, Republic of Korea. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Research Foundation of Korea, Chungbuk National University.
Abstract (Originalsprache)
As recent Multi-Layer Perceptron (MLP) mixer models have achieved state-of-the-art performance in time series forecasting, modeling each MLP-mixer as a separate expert within a mixture is expected to extend the representational capacity of the model, allowing each expert to be activated in response to time-varying inputs. However, extending MLP-mixers into a Mixture-of-Experts (MoE) architecture introduces a significant increase in the number of trainable parameters, rendering the model more challenging to train. To mitigate this problem, we propose a method that composes a fully trainable global expert and multiple non-trainable local experts. Specifically, our approach clones the weights of the global expert into the local experts and then modifies their weight distributions using moment learning, a recently proposed unconventional method for training neural networks. Concretely, each local expert is produced by applying moment-based transformations to a shared copy of the global expert’s weights, so that expert specialization is obtained without independently training the additional experts. Experimental results using a lightweight Time Series Mixer (TSMixer) architecture demonstrate that our method achieves performance competitive with fully trainable MoE counterparts, without introducing a significant increase in trainable parameters. Across multiple benchmark settings, the proposed model attains forecasting accuracy on par with, and in several cases favorable to, a fully trainable multi-expert baseline while adding only a small fraction of the extra trainable parameters that such a baseline requires, and this efficiency is further corroborated by measurements of memory footprint as well as an effect-size-based assessment of the observed differences.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Computer Science, Chungbuk National University, Cheongju 28644, Republic of Korea.
ORCID-Profile: Keon Myung Lee
Förderung
- National Research Foundation of Korea
- Chungbuk National University
Zitieren
APA
Hong, J., & Lee, K. M. (2026). Mixture of TSMixer Experts for Time Series Forecasting. Biomimetics, 11(6), 426. https://doi.org/10.3390/biomimetics11060426
Vancouver
Hong J, Lee KM. Mixture of TSMixer Experts for Time Series Forecasting. Biomimetics. 2026;11(6):426. doi:10.3390/biomimetics11060426
BibTeX
@article{Hong2026,
author = {Hong, Jaemoo and Lee, Keon Myung},
title = {Mixture of TSMixer Experts for Time Series Forecasting},
journal = {Biomimetics},
year = {2026},
volume = {11},
number = {6},
pages = {426},
doi = {10.3390/biomimetics11060426}
}
EUVIMED