Research map / Time series / Long-term forecasting
DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting
Siru Zhong, Yiqiu Liu, Zhiqing Cui, Zezhi Shao, Fei Wang, Qingsong Wen, Yuxuan Liang
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 DropoutTS to Informer | Informer · Synth-12 | MSE avg improvement: 46.0% avg MSE reduction | Improved |
| add DropoutTS to PatchTST, iTransformer, TimeMixer | strong backbones · Synth-12 | MSE: 0.7% to 2.8% | Improved |
| add DropoutTS to PatchTST on ETTm2 | PatchTST · ETTm2 | MSE: 0.280 -> 0.280 (0.0%) | No clear effect |
| remove global linear detrending (ablation) | DropoutTS · Synth-12 | MSE: 1.076 -> 1.642 without (41.7% worse than baseline) | Improved |
| add SFM anchor to threshold (ablation) | DropoutTS · Synth-12 | MSE: 1.520 -> 1.076 | Improved |
| add log-scale spectral normalization (ablation) | DropoutTS · Synth-12 | MSE: 1.308 -> 1.076 | Improved |
| use robust OLS detrend instead of simple detrend (ablation) | DropoutTS · Synth-12 | MSE: 1.574 -> 1.076 | Improved |
| initial sensitivity gamma 1.0 vs 5.0 vs 10.0 (ablation) | DropoutTS · Synth-12 | MSE | Mixed |
| add DropoutTS to PatchTST in few-shot ILI (10% data) | PatchTST · ILI few-shot | MAE: 1.333 -> 1.295 | Improved |
| add DropoutTS to PatchTST at 50% ILI data | PatchTST · ILI few-shot (50% data) | MSE: 3.143 -> 3.169 | Hurt |
| add DropoutTS to PatchTST for zero-shot transfer | PatchTST · ETT cross-dataset transfer | MAE: 0.402 -> 0.395 avg | Improved |
| hard low-pass denoising of input instead of DropoutTS (ablation) | PatchTST/TimeMixer · ILI | MAE: PatchTST 1.110 -> 1.162 | Hurt |
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.
- Informer — Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting (AAAI 2021) · better on 4 of 4 shared results
- TimesNet — TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis (ICLR 2023) · better on 2 of 2 shared results
- TimeMixer — TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (ICLR 2024) · better on 1 of 1 shared result
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
- Contextual Dropout: An Efficient Sample-Dependent Dropout Module (ICLR 2021 · similarity 0.86)
- Advanced Dropout: A Model-free Methodology for Bayesian Dropout Optimization (TPAMI 2021 · similarity 0.86)
- Stochastic Modified Equations and Dynamics of Dropout Algorithm (ICLR 2024 · similarity 0.85)
- SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series (NeurIPS 2025 · similarity 0.84)
- Multivariate Probabilistic Time Series Forecasting with Correlated Errors (NeurIPS 2024 · similarity 0.84)
- OmniCast: A Masked Latent Diffusion Model for Weather Forecasting Across Time Scales (NeurIPS 2025 · similarity 0.84)
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.
