mlsanityData to 14 Oct 2026

Task-Aware Mixture-of-Experts for Time Series Analysis

arXiv 2509.22279 ↗First posted 26 Sep 2025cs.LGcited by 2

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

Abstract (arXiv)

Time Series Analysis is widely used in various real-world applications such as weather forecasting, financial fraud detection, imputation for missing data in IoT systems, and classification for action recognization. Mixture-of-Experts (MoE), as a powerful architecture, though demonstrating effectiveness in NLP, still falls short in adapting to versatile tasks in time series analytics due to its task-agnostic router and the lack of capability in modeling channel correlations. In this study, we propose a novel, general MoE-based time series framework called PatchMoE to support the intricate ``knowledge'' utilization for distinct tasks, thus task-aware. Based on the observation that hierarchical representations often vary across tasks, e.g., forecasting vs. classification, we propose a Recurrent Noisy Gating to utilize the hierarchical information in routing, thus obtaining task-sepcific capability. And the routing strategy is operated on time series tokens in both temporal and channel dimensions, and encouraged by a meticulously designed Temporal & Channel Load Balancing Loss to model the intricate temporal and channel correlations. Comprehensive experiments on five downstream tasks demonstrate the state-of-the-art performance of PatchMoE.

Key facts. Cited by 2 papers 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 55 of its 74 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

55 of 74 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Moirai-MoE: Empowering Time Series Foundation Models with Sparse Mixture of Experts14 Oct
  2. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts24 Sep

Other references · 53

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  3. 2021Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsitycited by 3,073
  4. 2020Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecastingcited by 2,169
  5. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
  6. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,424
  7. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,210
  8. 2022TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysiscited by 1,058
  9. 2023iTransformer: Inverted Transformers Are Effective for Time Series Forecastingcited by 931
  10. 2021Scaling Vision with Sparse Mixture of Expertscited by 747
  11. 2022Mixture-of-Experts with Expert Choice Routingcited by 623
  12. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 614
38 more
  1. 2021CSDI: Conditional Score-based Diffusion Models for Probabilistic Time Series Imputationcited by 582
  2. 2021Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancycited by 394
  3. 2022Non-stationary Transformers: Exploring the Stationarity in Time Series Forecastingcited by 391
  4. 2016Expert Gate: Lifelong Learning with a Network of Expertscited by 372
  5. 2019ROCKET: Exceptionally fast and accurate time series classification using random convolutional kernelscited by 339
  6. 2018The UEA multivariate time series classification archive, 2018cited by 338
  7. 2024MACE: Mass Concept Erasure in Diffusion Modelscited by 307
  8. 2024TFB: Towards Comprehensive and Fair Benchmarking of Time Series Forecasting Methodscited by 249
  9. 2017Gate-Variants of Gated Recurrent Unit (GRU) Neural Networkscited by 242
  10. 2024Timer: Generative Pre-trained Transformers Are Large Time Series Modelscited by 208
  11. 2023FITS: Modeling Time Series with 10k Parameterscited by 195
  12. 2023TF-ICON: Diffusion-Based Training-Free Cross-Domain Image Compositioncited by 182
  13. 2025Sundial: A Family of Highly Capable Time Series Foundation Modelscited by 180
  14. 2022SAITS: Self-Attention-based Imputation for Time Seriescited by 169
  15. 2022Less Is More: Fast Multivariate Time Series Forecasting with Light Sampling-oriented MLP Structurescited by 160
  16. 2024TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysiscited by 116
  17. 2024Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advancescited by 110
  18. 2024CALF: Aligning LLMs for Time Series Forecasting via Cross-modal Fine-Tuningcited by 108
  19. 2024SparseTSF: Modeling Long-term Time Series Forecasting with 1k Parameterscited by 107
  20. 2024DUET: Dual Clustering Enhanced Multivariate Time Series Forecastingcited by 107
  21. 2024CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patternscited by 106
  22. 2024Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecastingcited by 104
  23. 2023DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detectioncited by 103
  24. 2024CATCH: Channel-Aware multivariate Time Series Anomaly Detection via Frequency Patchingcited by 94
  25. 2022Flowformer: Linearizing Transformers with Conservation Flowscited by 69
  26. 2025Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Conceptscited by 62
  27. 2025TAB: Unified Benchmarking of Time Series Anomaly Detection Methodscited by 60
  28. 2025A Comprehensive Survey of Deep Learning for Multivariate Time Series Forecasting: A Channel Strategy Perspectivecited by 58
  29. 2025Rethinking Irregular Time Series Forecasting: A Simple yet Effective Baselinecited by 55
  30. 2025ReAgent-V: A Reward-Driven Multi-Agent Framework for Video Understandingcited by 50
  31. 2024TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecastingcited by 40
  32. 2022Self-supervised Contrastive Representation Learning for Semi-supervised Time-Series Classificationcited by 39
  33. 2025DAG: A Dual Correlation Network for Time Series Forecasting with Exogenous Variablescited by 33
  34. 2024WcDT: World-centric Diffusion Transformer for Traffic Scene Generationcited by 32
  35. 2024EasyTime: Time Series Forecasting Made Easycited by 30
  36. 2024Hierarchical Classification Auxiliary Network for Time Series Forecastingcited by 26
  37. 2025LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimizationcited by 22
  38. 2024SSD-TS: Exploring the Potential of Linear State Space Models for Diffusion Models in Time Series Imputationcited by 18

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026TimeART: Towards Agentic Time Series Reasoning via Tool-Augmentationcited by 26
  2. 2026Position: Universal Time Series Foundation Models Rest on a Category Errorcited by 4

Similar, not cited

Close in wording, no citation either way

  1. 2025GateTS: Versatile and Efficient Forecasting via Attention-Inspired routed Mixture-of-Experts
  2. 2025Time Tracker: Mixture-of-Experts-Enhanced Foundation Time Series Forecasting Model with Decoupled Training Pipelines
  3. 2026Seg-MoE: Multi-Resolution Segment-wise Mixture-of-Experts for Time Series Forecasting Transformers
  4. 2026AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting
  5. 2025MoFE-Time: Mixture of Frequency Domain Experts for Time-Series Forecasting Models
  6. 2026WaveMoE: A Wavelet-Enhanced Mixture-of-Experts Foundation Model for Time Series Forecasting
  7. 2025Wavelet Mixture of Experts for Time Series Forecasting
  8. 2026Multi-Modal Time Series Prediction via Mixture of Modulated Experts