arXiv:2603.15158v3 Announce Type: replace 
Abstract: Addressing the domain adaptation problem becomes more challenging when distribution shifts across domains stem from latent confounders that affect both covariates and outcomes. Existing proxy-based approaches that address latent shift rely on a strong completeness assumption to uniquely determine (point-identify) a robust predictor. Completeness requires that proxies have sufficient information about variations in latent confounders. For imperfect proxies the mapping from confounders to the space of proxy distributions is non-injective, and multiple latent confounder values can generate the same proxy distribution. This breaks the completeness assumption and observed data are consistent with multiple potential predictors (set-identified). To address this, we introduce latent equivalent classes (LECs). LECs are defined as groups of latent confounders that induce the same conditional proxy distribution. We show that point-identification for the robust predictor remains achievable as long as multiple domains differ sufficiently in how they mix proxy-induced LECs to form the robust predictor. This domain diversity condition is formalized as a cross-domain rank condition on the mixture weights, which is substantially weaker assumption than completeness.
  We introduce the Proximal Quasi-Bayesian Active learning (PQAL) framework, which actively queries
  a small, targeted set of diverse domains that satisfy this rank condition. PQAL can recover the point-identified predictor, demonstrates robustness to varying degrees of shift and outperforms previous methods on synthetic data and semi-synthetic dSprites, IHDP, ACS Folktables datasets.

دراسة حديثة نُشرت على arXiv تتناول التحديات المتعلقة بتكيف المجال عندما تنشأ تحولات التوزيع من عوامل خفية تؤثر على كل من المتغيرات والنتائج. تقدم البحث مفهوم الفئات المعادلة الخفية (LECs) لتسهيل تحديد النقاط الخاصة بالمؤشرات القوية، حتى عندما تكون الوكلاء غير مثاليين، مما يكسر الافتراض التقليدي للاكتمال في الأساليب المعتمدة على الوكلاء الموجودة.

Un estudio reciente publicado en arXiv aborda los desafíos de la adaptación de dominio cuando los cambios de distribución surgen de confusores latentes que afectan tanto a las covariables como a los resultados. La investigación introduce el concepto de clases equivalentes latentes (LECs) para facilitar la identificación puntual de predictores robustos, incluso cuando los proxies son imperfectos, rompiendo así la suposición de completitud tradicional en los enfoques basados en proxies existentes.

Une étude récente publiée sur arXiv aborde les défis de l'adaptation de domaine lorsque les changements de distribution proviennent de confondants latents qui impactent à la fois les covariables et les résultats. La recherche introduit le concept de classes équivalentes latentes (LECs) pour faciliter l'identification ponctuelle de prédicteurs robustes, même lorsque les proxys sont imparfaits, rompant ainsi l'hypothèse de complétude traditionnelle dans les approches basées sur les proxys existants.

A recent study published on arXiv addresses the challenges of domain adaptation when distribution shifts arise from latent confounders that impact both covariates and outcomes. The research introduces the concept of latent equivalent classes (LECs) to facilitate point-identification of robust predictors, even when proxies are imperfect, thus breaking the traditional completeness assumption in existing proxy-based approaches.