UPLOTS: one diffusion model generates multi-dataset time series via scenario prompts
flosalim · x · 2026-09-24
UPLOTS tackles the fragmentation of synthetic time-series generation, where each new dataset or scenario traditionally requires training and maintaining a separate model.
- Approach: a single shared diffusion generator, guided by a pretrained transformer and controlled by scenario prompts, jointly trained across supported datasets.
- Usage: pick a supported dataset (traffic, energy load, etc.) and scenario (morning-peak, evening-peak, high-volatility) via the prompt; the same model samples from the corresponding learned distribution.
- Benefit: no per-dataset models, no retraining to switch scenarios — scenario choice becomes an explicit input to the generator.
This reduces engineering overhead for multi-domain time-series synthesis.
More from Research
- Open Replications Miss the Secret Sauce: HF Collection Curates Datasets for Jev-Style Models — vanstriendaniel · 2026-09-24
- The coin-flip test: why LLMs fail at calibration, which is the actual product — airesearch12 · 2026-09-24
- Experts forecast AI virology parity for 2030-2034; it arrived in April 2025 — davidmanheim · 2026-09-24
- In-house multi-vec retriever crushes API dense retrievers on memory search, ColBERT next? — bo_wangbo · 2026-09-24
- OmniVChat bilingual video-dialogue dataset trends on Hugging Face — Harland · 2026-09-24
- Uranus: diffusion-based robot simulator hits 24 FPS with open-ended rollout — D-Robotics · 2026-09-24