U$^3$CF: Unbiased, Unconfounding, and Unified Causal Framework for Multi-Target Domain Adaptation

Multi-target domain adaptation (MTDA) trains a model using a labeled source domain and several unlabeled target domains, aiming to enhance performance across all targets. However, existing methods lack a principled causal formulation and often rely on empirical domain-invariance enforcement, which can bias adaptation across targets. To fill this gap, we propose the **U**nbiased, **U**nconfounding, and **U**nified **C**ausal **F**ramework (**U$^3$CF**) for MTDA. To *unify* the alignment of multiple domains, we propose a prototype-driven alignment strategy that progressively updates prototypes by high-confidence target predictions, while the contrastive optimization objective jointly aligns target samples to semantic prototypes and preserves class discrimination. By formulating a structural causal model, we reveal that domain-invariant causal factors and domain-specific factors shape representations and labels, while the latter induce spurious label correlations across targets. Accordingly, U$^3$CF achieves *unbiased* prediction by disentangling representations into invariant causal components and domain-specific confounders and applying conditional intervention to *block confounding* effects while preserving invariant semantics. To ensure precise disentanglement, we leverage mutual information theory to derive a principled criterion for feature separation. Extensive experiments on four benchmarks demonstrate that U$^3$CF consistently outperforms leading methods.

Machine learning models often perform well on data similar to their training data, but their performance can drop when they are used in new environments. For example, an image recognition model trained on one type of data may not work equally well on images collected from different devices, backgrounds, or conditions. This paper studies multi-target domain adaptation, where a model is trained with labeled data from one source domain and unlabeled data from several target domains, so that it can work better across multiple new environments. We propose U$^3$CF, a causal learning framework that helps the model align different domains while reducing the influence of domain-specific biases. The method learns shared, task-relevant information across domains and separates it from domain-specific factors that may mislead prediction. By using causal intervention, the model makes more balanced predictions across target domains. Experiments on several benchmark datasets show that U$^3$CF improves performance compared with existing domain adaptation methods.