Bi-level Unbalanced Optimal Transport for Partial Domain Adaptation

2025/05/19 by Ziying Chen, Chen, Zi-Ying, Chuan-Xian Ren +3
Computer Science · #Anomaly Detection Techniques and Applications #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG) #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2506.08020

openalex publication_date 2025/05/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Partial domain adaptation (PDA) problem requires aligning cross-domain samples while distinguishing the outlier classes for accurate knowledge transfer. The widely used weighting framework tries to address the outlier classes by introducing the reweighed source domain with a similar label distribution to the target domain. However, the empirical modeling of weights can only characterize the sample-wise relations, which leads to insufficient exploration of cluster structures, and the weights could be sensitive to the inaccurate prediction and cause confusion on the outlier classes. To tackle these issues, we propose a Bi-level Unbalanced Optimal Transport (BUOT) model to simultaneously characterize the sample-wise and class-wise relations in a unified transport framework. Specifically, a cooperation mechanism between sample-level and class-level transport is introduced, where the sample-level transport provides essential structure information for the class-level knowledge transfer, while the class-level transport supplies discriminative information for the outlier identification. The bi-level transport plan provides guidance for the alignment process. By incorporating the label-aware transport cost, the local transport structure is ensured and a fast computation formulation is derived to improve the efficiency. Extensive experiments on benchmark datasets validate the competitiveness of BUOT.

Citations

• From Big to Small: Adaptive Learning to Partial-Set Domains
• Conditional Bures Metric for Domain Adaptation
• Select, Label, and Mix: Learning Discriminative Invariant Feature Representations for Partial Domain Adaptation
• A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation
• CO-Optimal Transport
• Domain Adversarial Reinforcement Learning for Partial Domain Adaptation
• Learning to Transfer Examples for Partial Domain Adaptation
• Contrastive Adaptation Network for Unsupervised Domain Adaptation
• Partial Adversarial Domain Adaptation
• Importance Weighted Adversarial Nets for Partial Domain Adaptation
• VisDA: The Visual Domain Adaptation Challenge
• Partial Transfer Learning with Selective Adversarial Networks
• Deep Hashing Network for Unsupervised Domain Adaptation
• Joint Distribution Optimal Transportation for Domain Adaptation
• Deep Residual Learning for Image Recognition
• Unbalanced optimal transport: Dynamic and Kantorovich formulations
• Adam: A Method for Stochastic Optimization
• Deep Domain Confusion: Maximizing for Domain Invariance
• ImageNet Large Scale Visual Recognition Challenge
• ImageNet Large Scale Visual Recognition Challenge
• Convergence of a Block Coordinate Descent Method for Nondifferentiable Minimization

Related

• Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport
• Optimal Transport for Domain Adaptation
• A Class-aware Optimal Transport Approach with Higher-Order Moment Matching for Unsupervised Domain Adaptation
• Self-Adaptive Partial Domain Adaptation
• Unified Optimal Transport Framework for Universal Domain Adaptation