mlsanityData to 7 Oct 2026

PatchMixer: A Patch-Mixing Architecture for Long-Term Time Series Forecasting

arXiv 2310.00655 ↗First posted 1 Oct 2023cs.LGcited by 20

Not on the map. Why?

Abstract (arXiv)

Although the Transformer has been the dominant architecture for time series forecasting tasks in recent years, a fundamental challenge remains: the permutation-invariant self-attention mechanism within Transformers leads to a loss of temporal information. To tackle these challenges, we propose PatchMixer, a novel CNN-based model. It introduces a permutation-variant convolutional structure to preserve temporal information. Diverging from conventional CNNs in this field, which often employ multiple scales or numerous branches, our method relies exclusively on depthwise separable convolutions. This allows us to extract both local features and global correlations using a single-scale architecture. Furthermore, we employ dual forecasting heads encompassing linear and nonlinear components to better model future curve trends and details. Our experimental results on seven time-series forecasting benchmarks indicate that compared with the state-of-the-art method and the best-performing CNN, PatchMixer yields 3.9% and 21.2% relative improvements, respectively, while being 2-3x faster than the most advanced method.

Key facts. Cited by 20 papers in the map (data through 7 Oct 2026). Builds on 25 of its 33 references that are in the map. First posted on arXiv 1 Oct 2023.

Builds on

25 of 33 references.

References · 25

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  3. 2014Going Deeper with Convolutionscited by 4,231
  4. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  5. 2016Gaussian Error Linear Units (GELUs)cited by 3,431
  6. 2021Efficiently Modeling Long Sequences with Structured State Spacescited by 2,424
  7. 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
  8. 2020Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecastingcited by 2,141
  9. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 1,981
  10. 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,654
  11. 2018An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modelingcited by 1,447
  12. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,400
13 more
  1. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,200
  2. 2016Instance Normalization: The Missing Ingredient for Fast Stylizationcited by 1,187
  3. 2022TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysiscited by 1,030
  4. 2017DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networkscited by 1,004
  5. 2017Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networkscited by 869
  6. 2023Hyena Hierarchy: Towards Larger Convolutional Language Modelscited by 388
  7. 2021SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interactioncited by 335
  8. 2023Long-term Forecasting with TiDE: Time-series Dense Encodercited by 327
  9. 2022Patches Are All You Need?cited by 201
  10. 2021CKConv: Continuous Kernel Convolution For Sequential Datacited by 117
  11. 2014Rigid-Motion Scattering for Texture Classificationcited by 52
  12. 2023MTS-Mixers: Multivariate Time Series Forecasting via Factorized Temporal and Channel Mixingcited by 52
  13. 2023PETformer: Long-term Time Series Forecasting via Placeholder-enhanced Transformercited by 22

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 5; 2025: 10; 2026: 5.

20 papers.

Sort:

By area:

  1. 2024CycleNet: Enhancing Time Series Forecasting through Modeling Periodic Patternscited by 102
  2. 2024PatchAD: A Lightweight Patch-based MLP-Mixer for Time Series Anomaly Detectioncited by 12
  3. 2024Conditional Denoising Meets Polynomial Modeling: A Flexible Decoupled Framework for Time Series Forecastingcited by 3
  4. 2025KANMixer: a minimal KAN-centered mixer for long-term time series forecastingcited by 3
  5. 2024A New Perspective on Time Series Anomaly Detection: Faster Patch-based Broad Learning Systemcited by 2
  6. 2025ModelRadar: Aspect-based Forecast Evaluationcited by 2
  7. 2025MFRS: A Multi-Frequency Reference Series Approach to Scalable and Accurate Time-Series Forecastingcited by 1
  8. 2025TriP-LLM: A Tri-Branch Patch-wise Large Language Model Framework for Time-Series Anomaly Detectioncited by 1
  9. 2025TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecastingcited by 1
  10. 2026PMDformer: Patch-Mean Decoupling Information Transformer for Long-term Forecastingcited by 1
  11. 2024IIP-Mixer:Intra-Inter Patch Mixing Architecture for Battery Remaining Useful Life Predictioncited by 0
  12. 2025Logo-LLM: Local and Global Modeling with Large Language Models for Time Series Forecastingcited by 0
8 more
  1. 2025JustDense: Just using Dense instead of Sequence Mixer for Time Series analysiscited by 0
  2. 2025Echo Flow Networkscited by 0
  3. 2025Naga: Vedic Encoding for Deep State Space Modelscited by 0
  4. 2025DB2-TransF: All You Need Is Learnable Daubechies Wavelets for Time Series Forecastingcited by 0
  5. 2026ITS-Mina: A Harris Hawks Optimization-Based All-MLP Framework with Iterative Refinement and External Attention for Multivariate Time Series Forecastingcited by 0
  6. 2026QuITE: Query-Based Irregular Time Series Embeddingcited by 0
  7. 2026Challenges of Explainability in Continual Learning for Time Series Forecastingcited by 0
  8. 2026Attention-based Experience Replay Framework for Continual Learning of Agnostic Time Series Forecasting Modelscited by 0