Evaluating Temporal Plasticity in Foundation Time Series Models for Incremental Fine-tuning

--Time series foundation models excel at diverse time series forecasting tasks, but their capacity for continuous improvement through incremental learning remains unexplored. We present the first comprehensive study investigating these models' temporal plasticity--their ability to progressively enhance performance through continual learning while maintaining existing capabilities. Through experiments on real-world datasets exhibiting distribution shifts, we evaluate both conventional deep learning models and foundation models using a novel continual learning framework. Our findings reveal that while traditional models struggle with performance deterioration during incremental fine-tuning, foundation models like Time-MoE and Chronos demonstrate sustained improvement in predictive accuracy. This suggests that optimizing foundation model fine-tuning strategies may be more valuable than developing domain-specific small models. Our research introduces new evaluation methodologies and insights for developing foundation time series models with robust continuous learning capabilities. Time series data is a fundamental modality that underpins dynamic systems and plays a crucial role across a wide range of real-world applications, from finance [1] and healthcare [2] to transportation [3] and environmental monitoring [4]. In various domains, the collection of time series data involves diverse types, each characterized by distinct time-varying temporal structures, inter-series correlations, and complex distributions. These unique properties have traditionally driven the development of domain-specific deep learning architectures, each tailored to address the specific challenges of different time series analysis tasks.