Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series

arXiv:2401.03955v4 Announce Type: replace
Abstract: Large pre-trained models for zero/few-shot learning excel in language and vision domains but encounter challenges in multivariate time series (TS) due to the diverse nature and scarcity of publicly available pre-training data. Consequently, there has been a recent surge in utilizing pre-trained large language models (LLMs) with token adaptations for TS forecasting. These approaches employ cross-domain transfer learning and surprisingly yield impressive results. However, these models are typically very slow and large (~billion parameters) and …

abstract arxiv challenges cs.ai cs.lg data diverse domains excel few-shot few-shot learning forecasting language multivariate nature pre-trained models pre-training series time series training training data type vision