Research map / Time series / Foundation models for time series
Temporal Source Recovery for Time-Series Source-Free Unsupervised Domain Adaptation
Yucheng Wang, Peiliang Gong, Min Wu, Felix Ott, Xiaoli Li, Lihua Xie, Zhenghua 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 |
|---|---|---|---|
| add recovery module generating a source-like domain (ablation) | TemSR baseline · UCI-HAR | MF1 (%): 84.99 -> 90.00 | Improved |
| add recovery module generating a source-like domain (ablation) | TemSR baseline · MFD | MF1 (%): 78.53 -> 92.46 | Improved |
| add recovery module generating a source-like domain (ablation) | TemSR baseline · Sleep-EDF SSC | MF1 (%): 60.16 -> 62.84 | Improved |
| add enhancement module (anchor-based diversity maximization) (ablation) | TemSR w/ Recov. · UCI-HAR | MF1 (%): 90.00 -> 92.10 | Improved |
| add enhancement module (anchor-based diversity maximization) (ablation) | TemSR w/ Recov. · Sleep-EDF SSC | MF1 (%): 62.84 -> 63.86 | Improved |
| add local context-aware regularization instead of sample-level entropy (ablation) | TemSR · UCI-HAR | MF1 (%): 90.72 -> 92.10 | Improved |
| add entropy consistency across contexts (ablation) | TemSR · UCI-HAR | MF1 (%): 91.50 -> 92.10 | Improved |
| add early context patch set (ablation) | TemSR · UCI-HAR | MF1 (%): 91.20 -> 92.10 | Improved |
| add late context patch set (ablation) | TemSR · UCI-HAR | MF1 (%): 91.16 -> 92.10 | Improved |
| add recovered context patch set (ablation) | TemSR · MFD | MF1 (%): 92.96 -> 93.24 | Improved |
| add complete (global) context alongside local contexts (ablation) | TemSR · UCI-HAR | MF1 (%): 91.04 -> 92.10 | Improved |
| add anchor guidance loss to diversity maximization (ablation) | TemSR · UCI-HAR | MF1 (%): 88.91 -> 92.10 | Improved |
Closest papers by meaning
- Population Aware Diffusion for Time Series Generation (AAAI 2025 · similarity 0.85)
- Enhancing Multivariate Time-Series Domain Adaptation via Contrastive Frequency Graph Discovery and Language-Guided Adversary Alignment (AAAI 2025 · similarity 0.85)
- Efficient Source-Free Time-Series Adaptation via Parameter Subspace Disentanglement (ICLR 2025 · similarity 0.84)
- TimeDP: Learning to Generate Multi-Domain Time Series with Domain Prompts (AAAI 2025 · similarity 0.83)
- Time-MMD: Multi-Domain Multimodal Dataset for Time Series Analysis (NeurIPS 2024 · similarity 0.83)
- TS-RAG: Retrieval-Augmented Generation based Time Series Foundation Models are Stronger Zero-Shot Forecaster (NeurIPS 2025 · similarity 0.83)
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.
