TimeDP
Summary
TimeDP is a multi-domain time-series diffusion model that learns semantic prototypes and uses prototype-assignment weights as domain prompts for few-shot generation.
Official Artifacts
- Paper: https://arxiv.org/abs/2501.05403
- Code: https://github.com/microsoft/TimeCraft/tree/main/TimeDP
- Microsoft Research blog: https://www.microsoft.com/en-us/research/articles/timedp-creating-cross-domain-synthetic-time-series-data/
Role In The Wiki
TimeDP anchors the prototype/domain-prompt branch of TimeCraft. It is relevant to synthetic data and cross-domain adaptation, but not to action-conditioned modeling.