Research map / Time series / Foundation models for time series
SwiftTS: A Swift Selection Framework for Time Series Pre-trained Models via Multi-task Meta-Learning
Tengxue Zhang, Biao Ouyang, Yang Shu, Xinyang Chen, Chenjuan Guo, Bin Yang
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 |
|---|---|---|---|
| use all three model embeddings (meta, topology, function) instead of any single (ablation) | SwiftTS model encoder · 14 datasets avg over horizons | weighted Kendall tau: 0.267-0.370 (single) -> 0.442 | Improved |
| add functional embedding v_c to meta+topology embeddings (ablation) | SwiftTS · 14 datasets avg | weighted Kendall tau: 0.369 -> 0.442 | Improved |
| add topological embedding to meta+functional embeddings (ablation) | SwiftTS · 14 datasets avg | weighted Kendall tau: 0.411 -> 0.442 | Improved |
| add transferable cross-task meta-learning (ablation) | SwiftTS · 14 datasets, 4 horizons | weighted Kendall tau | Improved |
| train with IID rather than OOD test datasets (ablation) | SwiftTS · 4 datasets, 4 horizons | weighted Kendall tau | Improved |
| replace feature-based selection with SwiftTS learned selection | model selection · 14 datasets | Pr(top1): 0.304 -> 0.339 | Improved |
| select via SwiftTS vs per-dataset winners | SwiftTS vs baselines · 14 datasets x 4 horizons | weighted Kendall tau top-1 count: 28 vs 8 (Etran) | Mixed |
| use SwiftTS instead of fine-tuning all models | selection cost · ETTh1 | runtime: 4.97 x 10^4 s fine-tuning vs 1,000-4,000 s selection | Improved |
Closest papers by meaning
- Large Pre-trained time series models for cross-domain Time series analysis tasks (NeurIPS 2024 · similarity 0.86)
- TSLANet: Rethinking Transformers for Time Series Representation Learning (ICML 2024 · similarity 0.85)
- UP2ME: Univariate Pre-training to Multivariate Fine-tuning as a General-purpose Framework for Multivariate Time Series Analysis (ICML 2024 · similarity 0.85)
- TStruct: Learning Shared Temporal Structures for Long-Term Time-Series Forecasting (NeurIPS 2026 · similarity 0.85)
- TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting (ICLR 2024 · similarity 0.85)
- SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series (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.
