Review: Time-Series Foundation Models�Pretraining Recipes and Risks
Abstract
We synthesize methods for pretraining large models on multivariate time series, covering masking objectives, frequency-domain augmentations, and risk factors including memorization and label leakage.
Cite this article
Eriksen, J. & Santos, H. (2025). Review: Time-Series Foundation Models�Pretraining Recipes and Risks. Research Explorations in Global Knowledge & Technology (REGKT), 3 (7). Retrieved from https://regkt.com/article.php?id=132&slug=review-time-series-foundation-models-pretraining-recipes-and-risks