Connecting Domains and Contrasting Samples: A Ladder for Domain Generalization

2025-10-19

AI Summary

Domain generalization (DG) aims to enhance model generalization to unseen target domains using only source-domain data; however, distribution shifts between training and testing often degrade performance. To address this, we propose Domain-Connected Contrastive Learning (DCCL), a novel framework comprising three key components: (1) constructing intra-class cross-domain positive pairs to strengthen class-wise connectivity across domains; (2) introducing a model anchoring mechanism to stabilize representation learning; and (3) jointly optimizing a generative transformation loss to enforce robust feature extraction. DCCL operates without domain labels or access to target-domain data. Extensive experiments on five standard DG benchmarks demonstrate consistent and significant improvements over state-of-the-art methods. The implementation is publicly available.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Representation Learning for Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Distribution shifts between training and testing samples frequently occur in practice and impede model generalization performance. This crucial challenge thereby motivates studies on domain generalization (DG), which aim to predict the label on unseen target domain data by solely using data from source domains. It is intuitive to conceive the class-separated representations learned in contrastive learning (CL) are able to improve DG, while the reality is quite the opposite: users observe directly applying CL deteriorates the performance. We analyze the phenomenon with the insights from CL theory and discover lack of intra-class connectivity in the DG setting causes the deficiency. We thus propose a new paradigm, domain-connecting contrastive learning (DCCL), to enhance the conceptual connectivity across domains and obtain generalizable representations for DG. On the data side, more aggressive data augmentation and cross-domain positive samples are introduced to improve intra-class connectivity. On the model side, to better embed the unseen test domains, we propose model anchoring to exploit the intra-class connectivity in pre-trained representations and complement the anchoring with generative transformation loss. Extensive experiments on five standard DG benchmarks are performed. The results verify that DCCL outperforms state-of-the-art baselines even without domain supervision. The detailed model implementation and the code are provided through https://github.com/weitianxin/DCCL

Problem

• Addressing performance degradation when applying contrastive learning to domain generalization
• Improving intra-class connectivity across domains for better generalization
• Developing domain-connecting contrastive learning without requiring domain supervision

Innovation

• Domain-connecting contrastive learning enhances cross-domain connectivity
• Aggressive data augmentation improves intra-class sample connectivity
• Model anchoring exploits pre-trained representations for generalization