Research map / Time series / Foundation models for time series
Towards a Unified Generative Model for Scarce Time Series with Domain Experts
Zihao Yao, Qi Zheng, Jiankai Zuo, Yaying Zhang
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 MoDE router (domain prompts and prototypes) to expert gating (ablation) | TimeMoDE · Few-shot generation, 10% | c-FID: 0.302 -> 0.240 | Improved |
| add MoDE router to expert gating (ablation) | TimeMoDE · Few-shot generation, 10% | Discriminative score: 0.102 -> 0.078 | Improved |
| add MoDE router to expert gating (ablation) | TimeMoDE · Few-shot generation, 25 samples | c-FID: 1.235 -> 1.029 | Improved |
| use MoDE instead of dense expert layer (ablation) | TimeMoDE · Few-shot generation, 10% | c-FID: 0.875 -> 0.240 | Improved |
| use MoDE instead of dense expert layer (ablation) | TimeMoDE · Few-shot generation, 10% | Discriminative score: 0.211 -> 0.078 | Improved |
| use MoDE instead of dense expert layer (ablation) | TimeMoDE · Few-shot generation, 25 samples | c-FID: 1.438 -> 1.029 | Improved |
| add pre-training on multi-domain data (ablation) | TimeMoDE · Few-shot generation, 10% | c-FID: 0.882 -> 0.240 | Improved |
| add pre-training on multi-domain data (ablation) | TimeMoDE · Few-shot generation, 10% | Discriminative score: 0.167 -> 0.078 | Improved |
| add pre-training on multi-domain data (ablation) | TimeMoDE · Few-shot generation, 25 samples | c-FID: 1.525 -> 1.029 | Improved |
| add diffusion timestep conditioning in experts (ablation) | TimeMoDE · Few-shot generation, 10% | c-FID: 0.244 -> 0.240 | No clear effect |
| add diffusion timestep conditioning in experts (ablation) | TimeMoDE · Few-shot generation, 10% | Discriminative score: 0.097 -> 0.078 | Improved |
| add diffusion timestep conditioning in experts (ablation) | TimeMoDE · Few-shot generation, 25 samples | Discriminative score: 0.198 -> 0.183 | 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.
- DiffusionTS — Diffusion-TS: Interpretable Diffusion for General Time Series Generation (ICLR 2024) · better on 3 of 3 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
- Time Series Generation Under Data Scarcity: A Unified Generative Modeling Approach (NeurIPS 2025 · similarity 0.92)
- TimeDP: Learning to Generate Multi-Domain Time Series with Domain Prompts (AAAI 2025 · similarity 0.91)
- TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting (ICLR 2024 · similarity 0.89)
- Diffusion-TS: Interpretable Diffusion for General Time Series Generation (ICLR 2024 · similarity 0.89)
- TimeDART: A Diffusion Autoregressive Transformer for Self-Supervised Time Series Representation (ICML 2025 · similarity 0.88)
- Towards a General Time Series Forecasting Model with Unified Representation and Adaptive Transfer (ICML 2025 · similarity 0.88)
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.
