Research map / Time series / Long-term forecasting
DSENet: A Novel Dual-Stream Enhancement Network for Multi-Scale Non-Stationary Time Series Forecasting
Yuhan Wang, Yuanyuan Zou, Jie Cheng, Bin Dai, Jinhong Guo
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 LoRE block to local stream (ablation) | DSENet · OhioT1DM (HL=2h, FL=2h) | error: 0.423 -> 0.394 | Improved |
| add forget gate to global Mamba (ablation) | DSENet · OhioT1DM (HL=2h, FL=2h) | error: 0.417 -> 0.394 | Improved |
| add dynamic multi-scale patching (DMSP) (ablation) | DSENet · DCLP5 (HL=2h, FL=2h) | error: 0.498 -> 0.460 | Improved |
| add DSSAF adaptive fusion (ablation) | DSENet · DCLP5 (HL=2h, FL=2h) | error: 0.475 -> 0.460 | Improved |
| add bidirectional modeling (ablation) | DSENet · OhioT1DM (HL=2h, FL=2h) | error: 0.395 -> 0.394 | No clear effect |
| use parallel dual-stream instead of serial Mambaformer-style stacking (ablation) | DSENet · DCLP5 (HL=2h, FL=2h) | error: 0.542 -> 0.460 | Improved |
| add dual-stream enhancement mechanism (DFB-Mamba + LoRE) (ablation) | DSENet · DCLP5 (HL=2h) | error: 0.495 -> 0.460 | Improved |
| add DFB-Mamba and LoRE together to reduce inter-stream coupling (ablation) | baseline dual-stream · inter-stream CKA | CKA: 0.567 -> 0.472 | Improved |
| add DFB-Mamba alone (ablation) | baseline dual-stream · inter-stream CKA | CKA: 0.567 -> 0.560 | No clear effect |
| use longer history (6 days vs 2 hours) | DSENet · OhioT1DM averaged MSE | MSE: 0.468 -> 0.488 | Hurt |
| add Gaussian noise sigma=0.05 to inputs | DSENet · robustness test | MSE: 0.516 -> 0.522 | No clear effect |
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.
- TimeFilter — TimeFilter: Patch-Specific Spatial-Temporal Graph Filtration for Time Series Forecasting (ICML 2025) · better on 4 of 4 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
- xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition (AAAI 2025 · similarity 0.85)
- GlucoFM-Bench: Benchmarking Time-Series Foundation Models for Blood Glucose Forecasting (NeurIPS 2026 · similarity 0.85)
- MSGNet: Learning Multi-Scale Inter-series Correlations for Multivariate Time Series Forecasting (AAAI 2024 · similarity 0.83)
- MPTSNet: Integrating Multiscale Periodic Local Patterns and Global Dependencies for Multivariate Time Series Classification (AAAI 2025 · similarity 0.83)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.83)
- TVNet: A Novel Time Series Analysis Method Based on Dynamic Convolution and 3D-Variation (ICLR 2025 · similarity 0.82)
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.
