Partial Invariance for Classification under Structural Causal Models
Abstract
Making predictions in an unseen environment from data collected across multiple training environments is challenging. This challenge is particularly pronounced when the target-generating mechanism varies across environments. In this work, we introduce environment-label separation, a form of partial environmental invariance that decomposes the target conditional distribution into an environment-invariant component and a label-invariant component. We show that this partial invariance enables estimation of the target conditional distribution in an unseen environment when using labeled training data together with unlabeled test features. Within the structural causal model framework, we provide sufficient conditions for environment-label separation. Building on these conditions, we develop an algorithm, dIMP-Classifier, which jointly minimizes an estimate of the test risk and encourages invariance across environments. We further provide finite-sample guarantees that characterize the test risk in terms of deviations from invariance, estimation error, and sample complexity. Experiments on simulated data and the real-world light tunnel Causal Chamber data demonstrate that the proposed approach can perform under interventions on the target mechanism and consistently outperforms the considered standard and invariance-based baselines.
Related
- Bimeromorphic invariance of ∂∂̅-property and blow-up formulae I: counterexamples and derived category approach
- Invariance of Clustering Operations in Causal Effect Identification
- Tracing the Loop: Non-Causal Computation, Partial Traces, & Postselected Entanglement
- Optimal Experiments for Partial Causal Effect Identification
- P-stationary structural equation modeling: Causal effect identification under temporally incomplete observations
BibTeX
@misc{goddard2026partial,
title = {{Partial Invariance for Classification under Structural Causal Models}},
author = {Austin Goddard and Yu Xiang},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.02254},
eprint = {2611.02254},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.02254},
}
RIS
TY - UNPB
TI - Partial Invariance for Classification under Structural Causal Models
AU - Austin Goddard
AU - Yu Xiang
PY - 2026
DA - 2026/11/01
PB - arXiv
DO - 10.48550/arxiv.2611.02254
UR - https://arxiv.org/abs/2611.02254
AB - Making predictions in an unseen environment from data collected across multiple training environments is challenging. This challenge is particularly pronounced when the target-generating mechanism varies across environments. In this work, we introduce environment-label separation, a form of partial environmental invariance that decomposes the target conditional distribution into an environment-invariant component and a label-invariant component. We show that this partial invariance enables estimation of the target conditional distribution in an unseen environment when using labeled training data together with unlabeled test features. Within the structural causal model framework, we provide sufficient conditions for environment-label separation. Building on these conditions, we develop an algorithm, dIMP-Classifier, which jointly minimizes an estimate of the test risk and encourages invariance across environments. We further provide finite-sample guarantees that characterize the test risk in terms of deviations from invariance, estimation error, and sample complexity. Experiments on simulated data and the real-world light tunnel Causal Chamber data demonstrate that the proposed approach can perform under interventions on the target mechanism and consistently outperforms the considered standard and invariance-based baselines.
ER -