Chaos Is a LADDER: Domain Generalization Beyond Invariance via Reweighting
Abstract
Domain generalization (DG) aims to learn from multiple source domains and generalize to unseen target domains. Most DG methods pursue invariance: they seek a causal representation whose prediction rule is invariant across domains. This principle is effective when the causal mechanism is stable, but becomes restrictive when the domain itself modulates how causal content maps to the response. In this case, directly feeding domain style into the predictor can create misleading shortcuts, since style does not by itself cause the response. Yet the apparent chaos of multiple styles can become a ladder: style can locate the unseen target domain among source domains and guide which domain-dependent prediction rules should be trusted. We propose \emph{Latent Adaptive Domain Disentanglement and Environment Reweighting} (LADDER), a fixed-model DG pipeline that learns causal/style representations, freezes the encoders, fits source-specific classifiers, and uses an unlabeled target-domain covariate set only at inference to compute weights over these fixed classifiers, with no target labels or model-state updates. We establish theoretical guarantees for source reweighting and validate LADDER on simulations, FMoW, and a location-grouped iWildCam protocol, with gains in overall and group-averaged accuracy.
Related
- ACOUSTIC DOMAIN SHIFT IN SPOKEN LANGUAGE IDENTIFICATION FROM SYSTEMATIC DOMAIN GENERALIZATION EVALUATION TO REAL-WORLD APPLICATION
- Rural Readiness: Domain Generalization with Adversarially Chosen Domains
- Linear statistics in the Coulomb gas: cumulants beyond rotational invariance
- Kriging Beyond Gaussianity: Fixed-Domain Asymptotic Optimality
- Coherent-State Integrals Beyond the Domain of Operator Products
BibTeX
@misc{jiang2026chaos,
title = {{Chaos Is a LADDER: Domain Generalization Beyond Invariance via Reweighting}},
author = {Yuhang Jiang and Fengchuan Zhang and Sanguo Zhang and Guojun Zhu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2607.26458},
eprint = {2607.26458},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2607.26458},
}
RIS
TY - UNPB
TI - Chaos Is a LADDER: Domain Generalization Beyond Invariance via Reweighting
AU - Yuhang Jiang
AU - Fengchuan Zhang
AU - Sanguo Zhang
AU - Guojun Zhu
PY - 2026
DA - 2026/09/22
PB - arXiv
DO - 10.48550/arxiv.2607.26458
UR - https://arxiv.org/abs/2607.26458
AB - Domain generalization (DG) aims to learn from multiple source domains and generalize to unseen target domains. Most DG methods pursue invariance: they seek a causal representation whose prediction rule is invariant across domains. This principle is effective when the causal mechanism is stable, but becomes restrictive when the domain itself modulates how causal content maps to the response. In this case, directly feeding domain style into the predictor can create misleading shortcuts, since style does not by itself cause the response. Yet the apparent chaos of multiple styles can become a ladder: style can locate the unseen target domain among source domains and guide which domain-dependent prediction rules should be trusted. We propose \emph{Latent Adaptive Domain Disentanglement and Environment Reweighting} (LADDER), a fixed-model DG pipeline that learns causal/style representations, freezes the encoders, fits source-specific classifiers, and uses an unlabeled target-domain covariate set only at inference to compute weights over these fixed classifiers, with no target labels or model-state updates. We establish theoretical guarantees for source reweighting and validate LADDER on simulations, FMoW, and a location-grouped iWildCam protocol, with gains in overall and group-averaged accuracy.
ER -