Research map / Time series / Long-term forecasting
ReNF: Rethinking the Design of Neural Long-Term Time Series Forecasters
Yihang Lu, Xianwei Meng, Enhong Chen
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 |
|---|---|---|---|
| increase number of sub-forecasts from K=1 to K=6 (ablation) | ReNF · Traffic (horizon 720) | MSE: 0.426 -> 0.402 | Improved |
| increase number of sub-forecasts from K=1 to K=6 (ablation) | ReNF · ETTm1 (horizon 720) | MSE: 0.411 -> 0.401 | Improved |
| increase number of sub-forecasts from K=1 to K=6 (ablation) | ReNF · Weather (horizon 720) | MSE: 0.311 -> 0.307 | Improved |
| add BDO recursive input concatenation and per-stage loss (versus plain Direct Output) (ablation) | ReNF · ETTh1 | MSE | Improved |
| add layers under BDO versus under DO (ablation) | ReNF · ETT/Weather/Traffic | MSE | Improved |
| add EMA weight averaging for evaluation (ablation) | ReNF · ETTh2 | valid-test loss consistency | Improved |
| tune EMA decay rate (ablation) | ReNF · Electricity | test loss | Mixed |
| oracle combination of sub-forecasts versus last forecast | ReNF · Traffic | MSE: 0.365 -> 0.254 | Improved |
| oracle combination of sub-forecasts versus last forecast | ReNF · ETTh1 | MSE: 0.391 -> 0.270 | Improved |
| use ReNF instead of TimeBridge | MLP forecaster · Traffic | MSE: 0.357 vs 0.365 | Hurt |
| use ReNF instead of TimeBridge for compute | MLP forecaster · Traffic FLOPs | FLOPs (GB): 479.231 -> 22.813 | 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.
- TimeBridge — TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecasting (ICML 2025) · 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
- ARM: Refining Multivariate Forecasting with Adaptive Temporal-Contextual Learning (ICLR 2024 · similarity 0.89)
- Unlocking the Power of Patch: Patch-Based MLP for Long-Term Time Series Forecasting (AAAI 2025 · similarity 0.88)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.88)
- Less is More: Unlocking Specialization of Time Series Foundation Models via Structured Pruning (NeurIPS 2025 · similarity 0.88)
- PGN: The RNN's New Successor is Effective for Long-Range Time Series Forecasting (NeurIPS 2024 · similarity 0.88)
- Unlocking the Power of LSTM for Long Term Time Series Forecasting (AAAI 2025 · 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.
