Research map / Time series / Long-term forecasting
Generalizing Multi-Scale Time-Series Modeling with a Single Operator
Cheonwoo Lee, Dooho Lee, Doyun Choi, Jaemin Yoo
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use the plain MLP predictor instead of TimeMixer trend-seasonal mixing (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.486 -> 0.480 | No clear effect |
| use per-element learnable scale parameters instead of one shared scale (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.489 -> 0.480 | Improved |
| use dataset-level scale parameters instead of sample-wise MLP-predicted scales (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.490 -> 0.480 | Improved |
| add the scaling mechanism rather than raw input only (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.492 -> 0.480 | Improved |
| use learnable distance-aware LDG scaling instead of moving average (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.493 -> 0.480 | Improved |
| use a normalized scaling-operator kernel instead of unnormalized convolution (ablation) | Sigma · ETTh1 predict-720 | MSE: 0.524 -> 0.480 | Improved |
| Sigma versus second-best model | Electricity long-term · Electricity | MSE: 0.185 -> 0.175 | Improved |
| Sigma versus second-best model | Traffic long-term · Traffic | MSE: 0.497 -> 0.458 | Improved |
| Sigma versus best baseline on low-dimensional data | Weather long-term · Weather | MSE: 0.245 -> 0.247 | No clear effect |
| Sigma on the small Others category | M4 Others · M4 Others | OWA: 1.014 -> 1.037 | Hurt |
| increase input length L | Sigma · ETTh1 predict-720 | MSE | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 8 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Adaptive Multi-Scale Decomposition Framework for Time Series Forecasting (AAAI 2025 · similarity 0.87)
- Scaling Law for Time Series Forecasting (NeurIPS 2024 · similarity 0.87)
- Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (ICLR 2025 · similarity 0.86)
- Multi-Scale Finetuning for Encoder-based Time Series Foundation Models (NeurIPS 2025 · similarity 0.86)
- Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization (ICML 2025 · similarity 0.86)
- Towards Neural Scaling Laws for Time Series Foundation Models (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
