TimeLAVA: Learning-Agnostic Data Valuation for Time Series

2026-06-17

AI Summary

This work addresses the challenge of data valuation in time series, where existing methods often fail due to their reliance on specific models or neglect of temporal dependencies, multi-scale structures, and non-stationary dynamics. The authors propose a model-agnostic framework that efficiently computes point-wise value scores by quantifying the marginal contribution of each time segment toward reducing the distributional discrepancy between evaluation and reference data—without requiring model training. The core innovation integrates selective wavelet transforms with unbalanced optimal transport to construct a multi-scale Wasserstein discrepancy measure robust to distribution shifts. Theoretical guarantees link the proposed valuation to generalization performance and provide sensitivity bounds against external outliers. Empirical results demonstrate significant improvements over state-of-the-art methods in anomaly detection, data pruning, and label noise identification tasks.

Abstract

Data valuation quantifies the intrinsic quality of individual samples to enable principled data curation, quality control, and robust learning. For time series in critical domains such as healthcare, finance, and industrial monitoring, effective valuation methods are essential yet fundamentally lacking. Existing approaches are either model-dependent, limiting their generalizability, or designed for i.i.d. data and thus fail to capture temporal dependencies, multi-scale patterns, and non-stationary dynamics inherent to sequential data. We introduce TimeLAVA, a learning-agnostic framework that values temporal segments by their marginal contribution to minimizing distributional discrepancy between evaluated and reference data. At its core is a novel Selective Wavelet-based Wasserstein discrepancy combining multi-scale wavelet transforms for temporal localization with unbalanced optimal transport for robustness to distributional shifts. Segment values are efficiently computed via sensitivity analysis without requiring model training and aggregated into point-wise scores. We provide theoretical guarantees linking valuation to model-agnostic generalization and prove bounded sensitivity to outlier contamination. Extensive experiments across anomaly detection, data pruning, and label noise detection demonstrate that TimeLAVA produces significantly more informative value scores than existing methods on diverse real-world datasets.

Problem

• time series
• data valuation
• learning-agnostic
• temporal dependencies
• non-stationary dynamics

Innovation

• learning-agnostic valuation
• time series
• wavelet transform
• optimal transport
• data curation