Meta-learning to Address Data Shift in Time Series Classification

TL;DR

This paper evaluates meta-learning techniques for time series classification under data shift, showing they adapt faster and more stably than traditional deep learning, especially in data-scarce scenarios, and introduces a new seismic benchmark.

Contribution

It systematically compares meta-learning and traditional deep learning methods for data shift in time-series classification and introduces SeisTask, a new benchmark for adaptive learning research.

Findings

Meta-learning achieves faster adaptation in data-scarce regimes.

Advantages of meta-learning diminish with more data and larger models.

Alignment between training and test distributions is crucial for performance.

Abstract

Across engineering and scientific domains, traditional deep learning (TDL) models perform well when training and test data share the same distribution. However, the dynamic nature of real-world data, broadly termed \textit{data shift}, renders TDL models prone to rapid performance degradation, requiring costly relabeling and inefficient retraining. Meta-learning, which enables models to adapt quickly to new data with few examples, offers a promising alternative for mitigating these challenges. Here, we systematically compare TDL with fine-tuning and optimization-based meta-learning algorithms to assess their ability to address data shift in time-series classification. We introduce a controlled, task-oriented seismic benchmark (SeisTask) and show that meta-learning typically achieves faster and more stable adaptation with reduced overfitting in data-scarce regimes and smaller model architectures. As data availability and model capacity increase, its advantages diminish, with TDL with fine-tuning performing comparably. Finally, we examine how task diversity influences meta-learning and find that alignment between training and test distributions, rather than diversity alone, drives performance gains. Overall, this work provides a systematic evaluation of when and why meta-learning outperforms TDL under data shift and contributes SeisTask as a benchmark for advancing adaptive learning research in time-series domains.