mlsanityData to 14 Oct 2026

WaveMoE: A Wavelet-Enhanced Mixture-of-Experts Foundation Model for Time Series Forecasting

arXiv 2604.10544 ↗First posted 12 Apr 2026cs.LG cs.AIcited by 1

On the map: Mixture-of-experts → MoE for time series and graphs

Abstract (arXiv)

Time series foundation models (TSFMs) have recently achieved remarkable success in universal forecasting by leveraging large-scale pretraining on diverse time series data. Complementing this progress, incorporating frequency-domain information yields promising performance in enhancing the modeling of complex temporal patterns, such as periodicity and localized high-frequency dynamics, which are prevalent in real-world time series. To advance this direction, we propose a new perspective that integrates explicit frequency-domain representations into scalable foundation models, and introduce WaveMoE, a wavelet-enhanced mixture-of-experts foundation model for time series forecasting. WaveMoE adopts a dual-path architecture that jointly processes time series tokens and wavelet tokens aligned along a unified temporal axis, and coordinates them through a shared expert routing mechanism that enables consistent expert specialization while efficiently scaling model capacity. Preliminary experimental results on 16 diverse benchmark datasets indicate that WaveMoE has the potential to further improve forecasting performance by incorporating wavelet-domain corpora.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in MoE for time series and graphs, one of 12 directions in Mixture-of-experts. Builds on 34 of its 46 references that are in the map. First posted on arXiv 12 Apr 2026.

Builds on

34 of 46 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025MoFE-Time: Mixture of Frequency Domain Experts for Time-Series Forecasting Models9 Jul
  2. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts24 Sep

Other references · 32

  1. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,073
  2. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
  3. 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,676
  4. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,210
  5. 2024Chronos: Learning the Language of Time Seriescited by 759
  6. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 753
  7. 2023Time-LLM: Time Series Forecasting by Reprogramming Large Language Modelscited by 700
  8. 2023A decoder-only foundation model for time-series forecastingcited by 692
  9. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 614
  10. 2023Large Language Models Are Zero-Shot Time Series Forecasterscited by 500
  11. 2024MOMENT: A Family of Open Time-series Foundation Modelscited by 482
  12. 2024Foundation Models for Time Series Analysis: A Tutorial and Surveycited by 239
20 more
  1. 2024GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluationcited by 212
  2. 2024Timer: Generative Pre-trained Transformers Are Large Time Series Modelscited by 208
  3. 2025Chronos-2: From Univariate to Universal Forecastingcited by 194
  4. 2025Sundial: A Family of Highly Capable Time Series Foundation Modelscited by 180
  5. 2025TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learningcited by 120
  6. 2025Moirai 2.0: When Less Is More for Time Series Forecastingcited by 69
  7. 2024Toto: Time Series Optimized Transformer for Observabilitycited by 48
  8. 2024A Comprehensive Survey of Deep Learning for Time Series Forecasting: Architectural Diversity and Open Challengescited by 47
  9. 2024Towards Neural Scaling Laws for Time Series Foundation Modelscited by 33
  10. 2024Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenizationcited by 27
  11. 2025Foundation Models for Spatio-Temporal Data Science: A Tutorial and Surveycited by 24
  12. 2025Output Scaling: YingLong-Delayed Chain of Thought in a Large Pretrained Time Series Forecasting Modelcited by 24
  13. 2019RobustTrend: A Huber Loss with a Combined First and Second Order Difference Regularization for Time Series Trend Filteringcited by 12
  14. 2025Towards Non-Stationary Time Series Forecasting with Temporal Stabilization and Frequency Differencingcited by 10
  15. 2025Multi-Order Wavelet Derivative Transform for Deep Time Series Forecastingcited by 4
  16. 2025TimeExpert: Boosting Long Time Series Forecasting with Temporal Mix of Expertscited by 4
  17. 2024Disentangling Imperfect: A Wavelet-Infused Multilevel Heterogeneous Network for Human Activity Recognition in Flawed Wearable Sensor Datacited by 3
  18. 2026A Unified Shape-Aware Foundation Model for Time Series Classificationcited by 3
  19. 2026OATS: Online Data Augmentation for Time Series Foundation Modelscited by 3
  20. 2025Bridging Distribution Gaps in Time Series Foundation Model Pretraining with Prototype-Guided Normalizationcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026TiRex-2: Generalizing TiRex to Multivariate Data and Streamingcited by 13

Similar, not cited

Close in wording, no citation either way

  1. 2025Time Tracker: Mixture-of-Experts-Enhanced Foundation Time Series Forecasting Model with Decoupled Training Pipelines
  2. 2026AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting
  3. 2025FreqMoE: Enhancing Time Series Forecasting through Frequency Decomposition Mixture of Experts
  4. 2025Wavelet Mixture of Experts for Time Series Forecasting
  5. 2026Speaking Numbers to LLMs: Multi-Wavelet Number Embeddings for Time Series Forecasting
  6. 2025Conversational Time Series Foundation Models: Towards Explainable and Effective Forecasting
  7. 2026TiMi: Empower Time Series Transformers with Multimodal Mixture of Experts
  8. 2025Small but Mighty: Enhancing Time Series Forecasting with Lightweight LLMs