Research map / Time series / Long-term forecasting
Sparse-Scale Transformer with Bidirectional Awareness for Time Series Forecasting
Ying Liu, Bo Liu, Sheng Huang, Gang Luo, Wenbo Hu, Meng Wang, Richang Hong
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 |
|---|---|---|---|
| add sparse scale selection (ablation) | SSformer · Weather (T=96) | MSE: 0.183 -> 0.156 | Improved |
| add sparse scale selection (ablation) | SSformer · Electricity (T=96) | MSE: 0.166 -> 0.141 | Improved |
| use convolutional tokenization instead of patching (ablation) | SSformer · Weather | MSE: 5.9% | Improved |
| add convolutional tokenization (ablation) | SSformer · Electricity (T=96) | MSE: 0.152 -> 0.141 | Improved |
| add coarse-to-fine interaction (ablation) | SSformer · Weather (T=96) | MSE: 0.162 -> 0.156 | Improved |
| add fine-to-coarse interaction (ablation) | SSformer · Weather (T=96) | MSE: 0.158 -> 0.156 | Improved |
| add coarse-to-fine interaction (ablation) | SSformer · Electricity (T=96) | MSE: 0.149 -> 0.141 | Improved |
| add fine-to-coarse interaction (ablation) | SSformer · Electricity (T=96) | MSE: 0.145 -> 0.141 | Improved |
| increase number of selected scales N from 1 to 3 (ablation) | SSformer · Weather | MSE | Improved |
| increase N to 4 or 5 (ablation) | SSformer · Weather | MSE | No clear effect |
| increase N to 4 or 5 (ablation) | SSformer · Electricity | MSE | Improved |
| replace multi-scale MTST fixed windows with sparse scales | SSformer vs MTST · Weather and Solar | MSE: 7.1% and 14.0% reduction | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- Times2D — Times2D: Multi-Period Decomposition and Derivative Mapping for General Time Series Forecasting (AAAI 2025) · better on 3 of 3 shared results
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 2 of 2 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 2 of 2 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
- SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention (ICML 2024 · similarity 0.90)
- A Lightweight Sparse Interaction Network for Time Series Forecasting (AAAI 2025 · similarity 0.89)
- Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting (ICLR 2024 · similarity 0.88)
- iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024 · similarity 0.88)
- Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (ICLR 2025 · similarity 0.88)
- TSLANet: Rethinking Transformers for Time Series Representation Learning (ICML 2024 · similarity 0.88)
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.
