Research map / Time series / Long-term forecasting
DeepBooTS: Dual-Stream Residual Boosting for Drift-Resilient Time-Series Forecasting
Daojun Liang, Jing Chen, Xiao Wang, Yinglong Wang, Shuo Li
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 |
|---|---|---|---|
| subtract (not add) block outputs in the input stream (ablation) | DeepBooTS · Weather | MSE: 0.262 -> 0.256 | Improved |
| add learnable gating coefficient to each block output (ablation) | DeepBooTS · Weather | MSE: 0.262 -> 0.249 | Improved |
| switch output-stream aggregation from addition to subtraction (ablation) | DeepBooTS · Traffic / Electricity / Weather | MSE | Improved |
| replace Full attention with Prob-Attention inside DeepBooTS (ablation) | DeepBooTS · Electricity | MSE: 0.311 -> 0.161 | Improved |
| replace Full attention with Prob-Attention inside DeepBooTS (ablation) | DeepBooTS · Weather | MSE: 0.634 -> 0.243 | Improved |
| deepen blocks from 4 to 8 in iTransformer vs DeepBooTS to 16 (ablation) | DeepBooTS · depth scaling | validation MSE | Improved |
| use DeepBooTS (input 96) instead of TimeMixer | TimeMixer · Weather (multivariate) | MSE: 0.245 vs 0.240 | Hurt |
| use DeepBooTS instead of Periodformer on univariate Exchange | Periodformer · Exchange (univariate) | MSE: 0.429 vs 0.353 | Hurt |
| use DeepBooTS* (input 336) instead of DeepBooTS (input 96) | DeepBooTS · PEMS (multivariate) | MSE: 0.109 -> 0.075 | Improved |
| use DeepBooTS instead of PSLD on large-scale datasets | PSLD · CBS and Milano | MSE: 8.9% and 6.2% reduction | 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.
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 6 of 6 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
- SIN: Selective and Interpretable Normalization for Long-Term Time Series Forecasting (ICML 2024 · similarity 0.87)
- SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series (NeurIPS 2025 · similarity 0.87)
- TimePFN: Effective Multivariate Time Series Forecasting with Synthetic Data (AAAI 2025 · similarity 0.87)
- Selective Learning for Deep Time Series Forecasting (NeurIPS 2025 · similarity 0.87)
- Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (ICLR 2025 · similarity 0.86)
- ProbTS: Benchmarking Point and Distributional Forecasting across Diverse Prediction Horizons (NeurIPS 2024 · similarity 0.86)
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.
