Search papers, labs, and topics across Lattice.
This paper introduces CoSPOT, a novel framework for online time series forecasting that utilizes a frozen pre-trained LLM and compositional spectral prompts to enhance adaptability in non-stationary environments. By decomposing time series data into frequency bases and composing prompts based on their amplitudes, CoSPOT effectively reduces the number of parameters updated during online adaptation while maintaining strong performance on unseen patterns. Experimental results across various real-world datasets reveal that CoSPOT outperforms existing methods, particularly in scenarios with significant distribution shifts and extended online phases.
CoSPOT achieves superior online time series forecasting by leveraging frequency-domain insights, allowing it to adapt to unseen patterns with minimal parameter updates.
To address the sequential and evolving nature of time series, the Online Time Series Forecasting (OTSF) task has been extensively studied in multiple domains. Existing research focuses on adapting to non-stationary environments by employing memory buffer-based retrieval strategies. However, we observe that such frameworks struggle with long-term adaptation and fail to generalize to unseen patterns. To this end, we introduce CoSPOT, an LLM-based online time series forecasting framework that leverages a pre-trained LLM as the backbone online forecaster, motivated by its strong few-shot capabilities. For efficient online adaptation, CoSPOT keeps the LLM frozen and employs compositional spectral prompts grounded in frequency-domain bases to guide the model with the overall distribution of the input, thereby substantially reducing the number of parameters updated during the online phase. Specifically, CoSPOT decomposes time series into frequency bases and composes the corresponding spectral basis prompts according to their amplitudes, allowing unseen patterns to be represented as new combinations of learned basis prompts. Our extensive experiments on real-world datasets demonstrate the superiority and practicality of CoSPOT across challenging online scenarios, including extended online phases and cross-dataset settings with substantial distribution shifts. Our code is available at https://github.com/seungyoon-Choi/CoSPOT.
