Pith. sign in

REVIEW 21 cited by

TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2405.14616 v1 pith:EOTMYFRJ submitted 2024-05-23 cs.LG cs.AI

TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting

Shiyu Wang , Haixu Wu , Xiaoming Shi , Tengge Hu , Huakun Luo , Lintao Ma , James Y. Zhang , Jun Zhou This is my paper
classification cs.LG cs.AI
keywords forecastingseriesmultiscaletimetimemixervariationsdecompositiondisentangled
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Time series forecasting is widely used in extensive applications, such as traffic planning and weather forecasting. However, real-world time series usually present intricate temporal variations, making forecasting extremely challenging. Going beyond the mainstream paradigms of plain decomposition and multiperiodicity analysis, we analyze temporal variations in a novel view of multiscale-mixing, which is based on an intuitive but important observation that time series present distinct patterns in different sampling scales. The microscopic and the macroscopic information are reflected in fine and coarse scales respectively, and thereby complex variations can be inherently disentangled. Based on this observation, we propose TimeMixer as a fully MLP-based architecture with Past-Decomposable-Mixing (PDM) and Future-Multipredictor-Mixing (FMM) blocks to take full advantage of disentangled multiscale series in both past extraction and future prediction phases. Concretely, PDM applies the decomposition to multiscale series and further mixes the decomposed seasonal and trend components in fine-to-coarse and coarse-to-fine directions separately, which successively aggregates the microscopic seasonal and macroscopic trend information. FMM further ensembles multiple predictors to utilize complementary forecasting capabilities in multiscale observations. Consequently, TimeMixer is able to achieve consistent state-of-the-art performances in both long-term and short-term forecasting tasks with favorable run-time efficiency.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 21 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. $\text{DT}^2$: Decision-Targeted Digital Twins

    cs.LG 2026-06 unverdicted novelty 7.0

    DT² trains digital twins to preserve pairwise policy rankings from fitted Q-evaluation on offline data rather than minimizing one-step transition errors, improving policy ranking and reducing decision regret.

  2. How Good Can Linear Models Be for Time-Series Forecasting?

    cs.LG 2026-06 conditional novelty 6.0

    Optimized Ridge regression with series-specific preprocessing beats prior linear forecasters and exceeds Transformer, MLP, and CNN baselines on six of eight time-series benchmarks.

  3. How Good Can Linear Models Be for Time-Series Forecasting?

    cs.LG 2026-06 unverdicted novelty 6.0

    Tuned Ridge regression with series-specific preprocessing beats prior linear forecasters and complex neural baselines on most dataset-horizon pairs across eight benchmarks.

  4. MetaPS: Adaptive Programmatic Strategy Selection for Market Agents

    cs.AI 2026-06 unverdicted novelty 6.0

    MetaPS trains models via simulation rollouts to select from programmatic strategy libraries for market agents, yielding better performance than fixed or direct LLM baselines across model sizes.

  5. CausalMoE: A Billion-Scale Multimodal Foundation Model for Granger Causal Discovery with Pattern-Routed Heterogeneous Experts

    cs.LG 2026-06 unverdicted novelty 6.0

    CausalMoE is a multimodal foundation model with pattern-routed heterogeneous experts and LLM/VLM integration that claims new SOTA performance on supervised and few-shot Granger causal discovery benchmarks.

  6. Stationarity-Aware Retrieval-Augmented Time Series Forecasting

    cs.LG 2026-06 unverdicted novelty 6.0

    SARAF is a new retrieval-augmented framework for time series forecasting that uses temporal similarity followed by stationarity-modulated diversity selection and aggregation to improve accuracy under non-stationarity.

  7. What If We Let Forecasting Forget? A Sparse Bottleneck for Cross-Variable Dependencies

    cs.LG 2026-05 unverdicted novelty 6.0

    MS-FLOW uses a capacity-limited sparse routing mechanism to model only critical inter-variable dependencies in time series data, achieving state-of-the-art accuracy on 12 benchmarks with fewer but more reliable connections.

  8. PRISM-CTG: A Foundation Model for Cardiotocography Analysis with Multi-View SSL

    cs.LG 2026-04 unverdicted novelty 6.0

    PRISM-CTG is the first large-scale foundation model for cardiotocography that uses multi-view self-supervised learning on unlabeled data to learn transferable representations, outperforming baselines on seven downstre...

  9. DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting

    cs.AI 2026-01 conditional novelty 6.0

    DropoutTS uses spectral reconstruction residuals to set per-sample dropout rates, reporting consistent robustness gains on six forecasting backbones without architectural changes.

  10. Kairos: Toward Adaptive and Parameter-Efficient Time Series Foundation Models

    cs.LG 2025-09 unverdicted novelty 6.0

    Kairos is a parameter-efficient time series foundation model using dynamic patching tokenizer, mixture-of-size encoding, and spectral-conditioned positional embeddings to improve zero-shot forecasting on heterogeneous data.

  11. Dynamic Relational Priming Improves Transformer in Multivariate Time Series

    cs.LG 2025-09 conditional novelty 6.0

    Prime attention modulates attention keys and values per channel-pair and reports improved MTS forecasting accuracy across several benchmarks.

  12. PIER: Physics-Informed Environmental Retrieval for Time-Series Modeling

    cs.LG 2026-07 conditional novelty 5.0

    PIER augments embedding-based retrieval for lake modeling with a physics-aware stream scored by local verifiers, improving water temperature and dissolved oxygen prediction across 356 lakes.

  13. Multi-Scale Convolution with Optimal Transport Attention Effect on Multivariate Time Series

    cs.LG 2026-07 conditional novelty 5.0

    MSC-OT fuses multi-scale convolution and Sinkhorn OT regularization into inverted attention, producing small but consistent MSE/MAE gains over iTransformer and several recent MTS baselines on five public datasets.

  14. Reviving Error Correction in Modern Deep Time-Series Forecasting

    cs.LG 2026-05 unverdicted novelty 5.0

    UEC-STD is an architecture-agnostic corrector that uses seasonal-trend decomposition to mitigate autoregressive error accumulation in deep forecasters and reports gains across 4 backbones and 10 datasets.

  15. Perceive, Route and Modulate: Dynamic Pattern Recalibration for Time Series Forecasting

    cs.LG 2026-05 unverdicted novelty 5.0

    Dynamic Pattern Recalibration (DPR) adds a perceive-route-modulate pipeline that generates time-aware modulation vectors to recalibrate hidden states in forecasting models, improving performance across architectures w...

  16. CombinationTS: A Modular Framework for Understanding Time-Series Forecasting Models

    cs.LG 2026-05 unverdicted novelty 5.0

    CombinationTS decomposes time-series models into modules and finds that good embeddings let simple identity encoders match complex ones, while input structural priors give better performance-stability trade-offs than ...

  17. CollideNet: Hierarchical Multi-scale Video Representation Learning with Disentanglement for Time-To-Collision Forecasting

    cs.CV 2026-04 unverdicted novelty 5.0

    CollideNet achieves state-of-the-art time-to-collision forecasting on three public datasets by combining multi-scale spatial aggregation with temporal disentanglement of trend and seasonality in a hierarchical transformer.

  18. QuantFlow: A Federated Mamba-Based Post-Transformer Foundation Model for Time-Series Forecasting

    cs.LG 2026-07 conditional novelty 4.5

    QuantFlow combines inverted embeddings, bidirectional Mamba decoders, quantile regression, TSMixup, and three-round FedAvg to report competitive one-step MSE on ETTm1/Weather and retain accuracy across 20 non-IID clients.

  19. QuantFlow: A Federated Mamba-Based Post-Transformer Foundation Model for Time-Series Forecasting

    cs.LG 2026-07 reject novelty 4.0

    QuantFlow combines inverted embedding, bidirectional Mamba, quantile heads, and federated averaging into a time-series forecaster, reporting strong R² on eight benchmarks but with internally inconsistent headline numbers.

  20. CASE-NET: Deep Spatio-Temporal Representation Learning via Causal Attention and Channel Recalibration for Multivariate Time Series Classification

    cs.LG 2026-05 unverdicted novelty 4.0

    CASE-NET combines a causal temporal encoder with adaptive channel recalibration and reports new state-of-the-art accuracy on four of six evaluated multivariate time series tasks.

  21. Time Series Forecasting Through the Lens of Dynamics

    cs.LG 2025-07 unverdicted novelty 4.0

    Proposes dynamics-based analysis of time series models showing partial dynamics learning and end-positioning as key to performance, plus a plug-and-play improvement method.