Eigen-Value: Efficient Domain-Robust Data Valuation via Eigenvalue-Based Approach

TL;DR

The paper introduces Eigen-Value (EV), a spectral method for data valuation that enhances out-of-distribution robustness using only in-distribution data, with low computational costs.

Contribution

EV provides a novel spectral approximation of domain discrepancy and estimates data point contributions efficiently, improving OOD robustness without extra training.

Findings

EV achieves better OOD robustness in real-world datasets.

EV maintains stable data value rankings across domain shifts.

EV is computationally lightweight and practical for large-scale applications.

Abstract

Data valuation has become central in the era of data-centric AI. It drives efficient training pipelines and enables objective pricing in data markets by assigning a numeric value to each data point. Most existing data valuation methods estimate the effect of removing individual data points by evaluating changes in model validation performance under in-distribution (ID) settings, as opposed to out-of-distribution (OOD) scenarios where data follow different patterns. Since ID and OOD data behave differently, data valuation methods based on ID loss often fail to generalize to OOD settings, particularly when the validation set contains no OOD data. Furthermore, although OOD-aware methods exist, they involve heavy computational costs, which hinder practical deployment. To address these challenges, we introduce \emph{Eigen-Value} (EV), a plug-and-play data valuation framework for OOD robustness that uses only an ID data subset, including during validation. EV provides a new spectral approximation of domain discrepancy, which is the gap of loss between ID and OOD using ratios of eigenvalues of ID data's covariance matrix. EV then estimates the marginal contribution of each data point to this discrepancy via perturbation theory, alleviating the computational burden. Subsequently, EV plugs into ID loss-based methods by adding an EV term without any additional training loop. We demonstrate that EV achieves improved OOD robustness and stable value rankings across real-world datasets, while remaining computationally lightweight. These results indicate that EV is practical for large-scale settings with domain shift, offering an efficient path to OOD-robust data valuation.