Source-free Domain Adaptation via Avatar Prototype Generation and Adaptation

2021/06/18 by Zhen Qiu, Qiu, Zhen, Yifan Zhang +11 · 13 citations
Biochemistry, Genetics and Molecular Biology · Computer Science · #Cancer-related molecular mechanisms research #Domain Adaptation and Few-Shot Learning #Multimodal Machine Learning Applications #cs.CV

paper · pdf · doi:10.48550/arxiv.2106.15326

Accepted by IJCAI 2021

arxiv created 2021/06/18 · arxiv updated 2021/06/30

Abstract

We study a practical domain adaptation task, called source-free unsupervised domain adaptation (UDA) problem, in which we cannot access source domain data due to data privacy issues but only a pre-trained source model and unlabeled target data are available. This task, however, is very difficult due to one key challenge: the lack of source data and target domain labels makes model adaptation very challenging. To address this, we propose to mine the hidden knowledge in the source model and exploit it to generate source avatar prototypes (i.e., representative features for each source class) as well as target pseudo labels for domain alignment. To this end, we propose a Contrastive Prototype Generation and Adaptation (CPGA) method. Specifically, CPGA consists of two stages: (1) prototype generation: by exploring the classification boundary information of the source model, we train a prototype generator to generate avatar prototypes via contrastive learning. (2) prototype adaptation: based on the generated source prototypes and target pseudo labels, we develop a new robust contrastive prototype adaptation strategy to align each pseudo-labeled target data to the corresponding source prototypes. Extensive experiments on three UDA benchmark datasets demonstrate the effectiveness and superiority of the proposed method.

Citations

• Deep Domain Confusion: Maximizing for Domain Invariance
• Conditional Image Synthesis With Auxiliary Classifier GANs
• VisDA: The Visual Domain Adaptation Challenge
• Unleashing the Power of Contrastive Self-Supervised Visual Models via Contrast-Regularized Fine-Tuning

Cited by

• Collaborative Learning with Multiple Foundation Models for Source-Free Domain Adaptation
• Diffusion-Driven Progressive Target Manipulation for Source-Free Domain Adaptation
• Vicinity-Guided Discriminative Latent Diffusion for Privacy-Preserving Domain Adaptation
• DAM: Dual Active Learning with Multimodal Foundation Model for Source-Free Domain Adaptation
• COLA: Context-aware Language-driven Test-time Adaptation
• SCoDA: Self-supervised Continual Domain Adaptation
• Adapt in the Wild: Test-Time Entropy Minimization with Sharpness and Feature Regularization
• Personalized Feature Translation for Expression Recognition: An Efficient Source-Free Domain Adaptation Method
• φ-Adapt: A Physics-Informed Adaptation Learning Approach to 2D Quantum Material Discovery
• Source-Free Domain Adaptation via Multi-view Contrastive Learning
• Temporal Restoration and Spatial Rewiring for Source-Free Multivariate Time Series Domain Adaptation
• The Unreasonable Effectiveness of Large Language-Vision Models for Source-free Video Domain Adaptation
• Co-STAR: Collaborative Curriculum Self-Training with Adaptive Regularization for Source-Free Video Domain Adaptation
• Self-Bootstrapping for Versatile Test-Time Adaptation

Related

• Imbalance-Agnostic Source-Free Domain Adaptation via Avatar Prototype Alignment
• BMD: A General Class-balanced Multicentric Dynamic Prototype Strategy for Source-free Domain Adaptation
• A Prototype-Oriented Framework for Unsupervised Domain Adaptation
• Generation, augmentation, and alignment: A pseudo-source domain based method for source-free domain adaptation
• AVATAR: Adversarial self-superVised domain Adaptation network for TARget domain