mlsanityData to 7 Oct 2026

Dish-TS: A General Paradigm for Alleviating Distribution Shift in Time Series Forecasting

arXiv 2302.14829 ↗First posted 22 Feb 2023cs.LG cs.AIcited by 91

Not on the map. Why?

Abstract (arXiv)

The distribution shift in Time Series Forecasting (TSF), indicating series distribution changes over time, largely hinders the performance of TSF models. Existing works towards distribution shift in time series are mostly limited in the quantification of distribution and, more importantly, overlook the potential shift between lookback and horizon windows. To address above challenges, we systematically summarize the distribution shift in TSF into two categories. Regarding lookback windows as input-space and horizon windows as output-space, there exist (i) intra-space shift, that the distribution within the input-space keeps shifted over time, and (ii) inter-space shift, that the distribution is shifted between input-space and output-space. Then we introduce, Dish-TS, a general neural paradigm for alleviating distribution shift in TSF. Specifically, for better distribution estimation, we propose the coefficient net (CONET), which can be any neural architectures, to map input sequences into learnable distribution coefficients. To relieve intra-space and inter-space shift, we organize Dish-TS as a Dual-CONET framework to separately learn the distribution of input- and output-space, which naturally captures the distribution difference of two spaces. In addition, we introduce a more effective training strategy for intractable CONET learning. Finally, we conduct extensive experiments on several datasets coupled with different state-of-the-art forecasting models. Experimental results show Dish-TS consistently boosts them with a more than 20% average improvement. Code is available.

Key facts. Cited by 91 papers in the map (data through 7 Oct 2026). Builds on 13 of its 33 references that are in the map. First posted on arXiv 22 Feb 2023.

Builds on

13 of 33 references.

References · 13

  1. 2017Attention Is All You Needcited by 59,932
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  4. 2020Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecastingcited by 2,141
  5. 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,654
  6. 2017DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networkscited by 1,004
  7. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 744
  8. 2021Spectral Temporal Graph Neural Network for Multivariate Time-series Forecastingcited by 298
  9. 2021AdaRNN: Adaptive Learning and Forecasting of Time Seriescited by 120
  10. 2019Deep Factors for Forecastingcited by 86
  11. 2020Joint Air Quality and Weather Prediction Based on Multi-Adversarial Spatiotemporal Networkscited by 39
  12. 2022DEPTS: Deep Expansion Learning for Periodic Time Series Forecastingcited by 28
1 more
  1. 2022DSLOB: A Synthetic Limit Order Book Dataset for Benchmarking Forecasting Algorithms under Distributional Shiftcited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 10; 2024: 29; 2025: 24; 2026: 28.

91 papers.

Sort:

By area:

  1. 2023TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecastingcited by 170
  2. 2023Frequency-domain MLPs are More Effective Learners in Time Series Forecastingcited by 170
  3. 2023FourierGNN: Rethinking Multivariate Time Series Forecasting from a Pure Graph Perspectivecited by 126
  4. 2024DUET: Dual Clustering Enhanced Multivariate Time Series Forecastingcited by 104
  5. 2024Frequency Adaptive Normalization For Non-stationary Time Series Forecastingcited by 45
  6. 2024TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecastingcited by 38
  7. 2024Rethinking Channel Dependence for Multivariate Time Series Forecasting: Learning from Leading Indicatorscited by 32
  8. 2024Towards Neural Scaling Laws for Time Series Foundation Modelscited by 30
  9. 2024OpenCity: Open Spatio-Temporal Foundation Models for Traffic Predictioncited by 27
  10. 2024Time-Series Forecasting for Out-of-Distribution Generalization Using Invariant Learningcited by 26
  11. 2025Non-stationary Diffusion For Probabilistic Time Series Forecastingcited by 24
  12. 2023Calibration of Time-Series Forecasting: Detecting and Adapting Context-Driven Distribution Shiftcited by 23
38 more
  1. 2024Learning Pattern-Specific Experts for Time Series Forecasting Under Patch-level Distribution Shiftcited by 23
  2. 2025Bridging Past and Future: Distribution-Aware Alignment for Time Series Forecastingcited by 19
  3. 2024Parsimony or Capability? Decomposition Delivers Both in Long-term Time Series Forecastingcited by 18
  4. 2024Proactive Model Adaptation Against Concept Drift for Online Time Series Forecastingcited by 18
  5. 2023Estimating Treatment Effects from Irregular Time Series Observations with Hidden Confounderscited by 17
  6. 2025Out-of-Distribution Generalization in Time Series: A Surveycited by 15
  7. 2025Improving Time Series Forecasting via Instance-aware Post-hoc Revisioncited by 14
  8. 2024Deep Frequency Derivative Learning for Non-stationary Time Series Forecastingcited by 13
  9. 2024Rethinking the Power of Timestamps for Robust Time Series Forecasting: A Global-Local Fusion Perspectivecited by 13
  10. 2023Revisiting Long-term Time Series Forecasting: An Investigation on Linear Mappingcited by 11
  11. 2024Multi-Modality Spatio-Temporal Forecasting via Self-Supervised Learningcited by 9
  12. 2023Performative Time-Series Forecastingcited by 8
  13. 2025Selective Learning for Deep Time Series Forecastingcited by 8
  14. 2023MemDA: Forecasting Urban Time Series with Memory-based Drift Adaptationcited by 6
  15. 2024IN-Flow: Instance Normalization Flow for Non-stationary Time Series Forecastingcited by 6
  16. 2024Sparse-VQ Transformer: An FFN-Free Framework with Vector Quantization for Enhanced Time Series Forecastingcited by 6
  17. 2025TimeRecipe: A Time-Series Forecasting Recipe via Benchmarking Module Level Effectivenesscited by 6
  18. 2024Efficient Source-Free Time-Series Adaptation via Parameter Subspace Disentanglementcited by 4
  19. 2025Online time series prediction using feature adjustmentcited by 4
  20. 2026Towards Principled Test-Time Adaptation for Time Series Forecastingcited by 4
  21. 2024MARLP: Time-series Forecasting Control for Agricultural Managed Aquifer Rechargecited by 3
  22. 2025DELPHYNE: A Pre-Trained Model for General and Financial Time Seriescited by 3
  23. 2026TimeRFT: Stimulating Generalizable Time Series Forecasting for TSFMs via Reinforcement Finetuningcited by 3
  24. 2024Enhancing Spatio-temporal Quantile Forecasting with Curriculum Learning: Lessons Learnedcited by 2
  25. 2024D2Vformer: A Flexible Time Series Prediction Model Based on Time Position Embeddingcited by 2
  26. 2024Evolving Multi-Scale Normalization for Time Series Forecasting under Distribution Shiftscited by 2
  27. 2024FredNormer: Frequency Domain Normalization for Non-stationary Time Series Forecastingcited by 2
  28. 2024GAS-Norm: Score-Driven Adaptive Normalization for Non-Stationary Time Series Forecasting in Deep Learningcited by 2
  29. 2025Bridging Distribution Gaps in Time Series Foundation Model Pretraining with Prototype-Guided Normalizationcited by 2
  30. 2025Non-Stationary Time Series Forecasting Based on Fourier Analysis and Cross Attention Mechanismcited by 2
  31. 2025MetaEformer: Unveiling and Leveraging Meta-patterns for Complex and Dynamic Systems Load Forecastingcited by 2
  32. 2025EventTSF: Event-Aware Non-Stationary Time Series Forecastingcited by 2
  33. 2026T1: One-to-One Channel-Head Binding for Multivariate Time-Series Imputationcited by 2
  34. 2026On the Role of Reversible Instance Normalizationcited by 2
  35. 2024Robust Multivariate Time Series Forecasting against Intra- and Inter-Series Transitional Shiftcited by 1
  36. 2024Wormhole: Concept-Aware Deep Representation Learning for Co-Evolving Sequencescited by 1
  37. 2024Domain Fusion Controllable Generalization for Cross-Domain Time Series Forecasting from Multi-Domain Integrated Distributioncited by 1
  38. 2024Memory-enhanced Invariant Prompt Learning for Urban Flow Prediction under Distribution Shiftscited by 1

50 most cited of 91 All 91 (JSON)