Research map / Time series / Causal and generative time series
OneNet: Enhancing Time Series Forecasting Models under Concept Drift by Online Ensembling
YiFan Zhang, Qingsong Wen, Xue Wang, Weiqi Chen, Liang Sun, Zhang Zhang, Liang Wang, Rong Jin, Tieniu Tan
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 |
|---|---|---|---|
| Ensemble cross-time and cross-variable forecasters with the OCP block | OneNet (FSNet and Time-FSNet) · ETTH2, ETTm1, WTH, ECL | average MSE: 1.594 -> 0.747 | Improved |
| Ensemble TCN and Time-TCN | OneNet-TCN · ETTH2, ETTm1, WTH, ECL | average MSE: 1.549 -> 1.253 | Improved |
| Use the lightweight concatenated-feature variant instead of full OneNet (ablation) | OneNet · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 0.961 | Hurt |
| Replace OCP with simple averaging (ablation) | FSNet and Time-FSNet ensemble · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 0.918 | Hurt |
| Replace OCP with gating weights learned from expert predictions (ablation) | FSNet and Time-FSNet ensemble · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 0.820 | Hurt |
| Replace OCP with input-conditioned mixture-of-experts weights (ablation) | FSNet and Time-FSNet ensemble · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 1.005 | Hurt |
| Replace OCP with linear-regression weights (ablation) | FSNet and Time-FSNet ensemble · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 1.003 | Hurt |
| Remove the RL short-term weight, retaining EGD (ablation) | OneNet OCP block · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 0.797 | Hurt |
| Remove historical expert performance and learn weights from predictions and outcomes only (ablation) | OneNet OCP block · ETTH2, ETTm1, WTH, ECL | average MSE: 0.747 -> 0.778 | Hurt |
| Remove seasonal-trend decomposition (ablation) | Online-adapted PatchTST · ETTH2, ETTm1, WTH, ECL | average MSE: 1.512 -> 1.747 | Mixed |
| Remove instance normalization (ablation) | Online-adapted PatchTST · ETTH2, ETTm1, WTH, ECL | MSE: 1.510 -> 1.687 | Mixed |
| Disable online adaptation (ablation) | PatchTST · ETTH2, ETTm1, WTH, ECL | average MSE: 1.512 -> 1.619 | Hurt |
Later papers that compared with this method
3 later accepted papers report OneNet as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Disentangling Long-Short Term State Under Unknown Interventions for Online Time Series Forecasting (AAAI 2025 · better on 11 of 11 shared results)
- Lightweight Online Adaption for Time Series Foundation Model Forecasts (ICML 2025 · better on 9 of 9 shared results)
- Battling the Non-stationarity in Time Series Forecasting via Test-time Adaptation (AAAI 2025 · better on 1 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
- Online Time Series Forecasting with Theoretical Guarantees (NeurIPS 2025 · similarity 0.89)
- LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization (ICML 2025 · similarity 0.87)
- Disentangling Long-Short Term State Under Unknown Interventions for Online Time Series Forecasting (AAAI 2025 · similarity 0.87)
- Rethinking Post-Training Recipes for Multimodal Time-Series Forecasting (NeurIPS 2026 · similarity 0.85)
- Improving Time Series Forecasting via Instance-aware Post-hoc Revision (NeurIPS 2025 · similarity 0.85)
- xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories (NeurIPS 2025 · similarity 0.85)
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.
