Mamba4Cast: Efficient Zero-Shot Time Series Forecasting with State Space Models
This addresses the need for efficient and scalable forecasting models in domains like finance or weather, though it is incremental as it builds on existing Mamba and PFN architectures.
The paper tackles the problem of zero-shot time series forecasting by introducing Mamba4Cast, a model that generalizes robustly across diverse tasks without fine-tuning, achieving competitive performance with lower inference times than transformer-based models and scaling better with prediction length.
This paper introduces Mamba4Cast, a zero-shot foundation model for time series forecasting. Based on the Mamba architecture and inspired by Prior-data Fitted Networks (PFNs), Mamba4Cast generalizes robustly across diverse time series tasks without the need for dataset specific fine-tuning. Mamba4Cast's key innovation lies in its ability to achieve strong zero-shot performance on real-world datasets while having much lower inference times than time series foundation models based on the transformer architecture. Trained solely on synthetic data, the model generates forecasts for entire horizons in a single pass, outpacing traditional auto-regressive approaches. Our experiments show that Mamba4Cast performs competitively against other state-of-the-art foundation models in various data sets while scaling significantly better with the prediction length. The source code can be accessed at https://github.com/automl/Mamba4Cast.