PAITS: Pretraining and Augmentation for Irregularly-Sampled Time Series
This addresses the challenge of handling sparse and irregular time series data for applications in sequential human behavior analysis, though it appears incremental as it builds on existing pretraining and augmentation ideas.
The paper tackles the problem of pretraining and augmentation for irregularly-sampled time series, which are common in real-world data but not well-handled by existing methods, and presents PAITS, a framework that identifies effective strategies to improve pretraining across multiple datasets and domains.
Real-world time series data that commonly reflect sequential human behavior are often uniquely irregularly sampled and sparse, with highly nonuniform sampling over time and entities. Yet, commonly-used pretraining and augmentation methods for time series are not specifically designed for such scenarios. In this paper, we present PAITS (Pretraining and Augmentation for Irregularly-sampled Time Series), a framework for identifying suitable pretraining strategies for sparse and irregularly sampled time series datasets. PAITS leverages a novel combination of NLP-inspired pretraining tasks and augmentations, and a random search to identify an effective strategy for a given dataset. We demonstrate that different datasets benefit from different pretraining choices. Compared with prior methods, our approach is better able to consistently improve pretraining across multiple datasets and domains. Our code is available at \url{https://github.com/google-research/google-research/tree/master/irregular_timeseries_pretraining}.