ShapeCond: Fast Shapelet-Guided Dataset Condensation for Time Series Classification
This addresses storage and computation challenges in domains like finance and climate science by providing an efficient condensation method for time series data, though it is incremental as it builds on existing condensation techniques with a time-series-specific adaptation.
The paper tackles the problem of dataset condensation for time series classification by proposing ShapeCond, a framework that uses shapelet-guided optimization to preserve local temporal patterns, resulting in up to 29x faster synthesis over prior methods and improved downstream accuracy.
Time series data supports many domains (e.g., finance and climate science), but its rapid growth strains storage and computation. Dataset condensation can alleviate this by synthesizing a compact training set that preserves key information. Yet most condensation methods are image-centric and often fail on time series because they miss time-series-specific temporal structure, especially local discriminative motifs such as shapelets. In this work, we propose ShapeCond, a novel and efficient condensation framework for time series classification that leverages shapelet-based dataset knowledge via a shapelet-guided optimization strategy. Our shapelet-assisted synthesis cost is independent of sequence length: longer series yield larger speedups in synthesis (e.g., 29$\times$ faster over prior state-of-the-art method CondTSC for time-series condensation, and up to 10,000$\times$ over naively using shapelets on the Sleep dataset with 3,000 timesteps). By explicitly preserving critical local patterns, ShapeCond improves downstream accuracy and consistently outperforms all prior state-of-the-art time series dataset condensation methods across extensive experiments. Code is available at https://github.com/lunaaa95/ShapeCond.