Time-series foundation model – Definition & Meaning

Definition of time-series foundation model

: a foundation model pretrained on large and varied collections of time-ordered data so it can perform forecasting, classification, anomaly detection, or other time-series tasks across domains with little or no task-specific training
- The utility company tested a time-series foundation model on demand data from buildings it had never seen during training.
- A TSFM can provide a useful forecasting baseline before engineers build a model for one specific sensor network.

Origin & history

The expression extends foundation model to time-series analysis. It became established during the early 2020s as researchers pretrained reusable models on large collections of temporal data; systems such as TimesFM, Chronos, Moirai, and MOMENT helped make the category widely recognizable.

Test yourself

Which of these is the meaning of time-series foundation model?

Related terms

foundation model a large model trained on broad data that can be adapted to many different tasks, serving as… transformer a neural network design that processes a whole sequence at once and uses attention to work out… zero-shot done without giving the model any examples, relying only on instructions few-shot done by including a handful of examples in the prompt so the model can copy the pattern… distribution shift a difference between the statistical distribution of the data used to train or evaluate a… tabular foundation model a foundation model pretrained across many tables or synthetic data-generating tasks so that it…

Cite this entry

"Time-series foundation model." AI Dictionary, Dadgogo, https://dadgogo.com/dictionary/time-series-foundation-model/. Accessed 12 Oct. 2026.