FlowState: Sampling-Rate-Equivariant Time-Series Forecasting

FlowState is a new time series foundation model (TSFM) that combines a state space model (SSM) encoder with a functional basis decoder (FBD) to achieve sampling-rate-equivariant forecasting. This architecture allows for continuous-time modeling and dynamic time-scale adjustments, enabling generalization across various temporal resolutions without the need for retraining. FlowState has shown state-of-the-art performance on the GIFT-Eval benchmark while being one of the smallest TSFMs, making it a significant advancement for practitioners focused on efficient and adaptable time-series forecasting.