A pre-trained diffusion model with dynamic channel adaptation and dataset tokens generates time series from a few examples and outperforms from-scratch baselines on a 12-dataset few-shot benchmark.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Time Series Generation Under Data Scarcity: A Unified Generative Modeling Approach
A pre-trained diffusion model with dynamic channel adaptation and dataset tokens generates time series from a few examples and outperforms from-scratch baselines on a 12-dataset few-shot benchmark.