General Quantification of Covariate and Concept Shifts
Abstract
Generalization under distribution shift remains a core challenge in modern machine learning, yet existing learning bound theory is limited to narrow, idealized settings and is non-estimable from samples. In this paper, we bridge the gap between theory and practical applications. We first show that existing definition of concept shift breaks when the source and target supports mismatch. Leveraging entropic optimal transport, we propose a key notion: γ*-concept shifts, and derive a general error bound unifying covariate and γ*-concept shifts, which applies to broad loss functions, label spaces, and stochastic labeling. We further develop estimators for these shifts with concentration guarantees, and the DataShifts algorithm, which can quantify distribution shifts and estimate the error bound in most applications - a rigorous and general tool for analyzing learning error under distribution shift.
Related
- On the value distribution of the Riemann zeta-function with general shifts
- Dissecting microbial community structure and heterogeneity via multivariate covariate-adjusted clustering
- Automated, Efficient, and Model-Free Inference for Randomized Clinical Trials via Data-Driven Covariate Adjustment
- Convergence-Aware Pareto Selection of Covariate Scaling Transformations for Markov Deterioration Hazard Models: Evidence from Bridge Inspection Data
- Ensembles of Random Partition-Based Models under Covariate Shift
BibTeX
@misc{chen2026general,
title = {{General Quantification of Covariate and Concept Shifts}},
author = {Hongbo Chen and Li Charlie Xia},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.11918},
eprint = {2609.11918},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.11918},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - General Quantification of Covariate and Concept Shifts
AU - Hongbo Chen
AU - Li Charlie Xia
PY - 2026
DA - 2026/09/10
PB - arXiv
DO - 10.48550/arxiv.2609.11918
UR - https://arxiv.org/abs/2609.11918
AB - Generalization under distribution shift remains a core challenge in modern machine learning, yet existing learning bound theory is limited to narrow, idealized settings and is non-estimable from samples. In this paper, we bridge the gap between theory and practical applications. We first show that existing definition of concept shift breaks when the source and target supports mismatch. Leveraging entropic optimal transport, we propose a key notion: γ^*-concept shifts, and derive a general error bound unifying covariate and γ^*-concept shifts, which applies to broad loss functions, label spaces, and stochastic labeling. We further develop estimators for these shifts with concentration guarantees, and the DataShifts algorithm, which can quantify distribution shifts and estimate the error bound in most applications - a rigorous and general tool for analyzing learning error under distribution shift.
ER -