Model Adaptation: Unsupervised Domain Adaptation without Source Data

Abstract

In this paper, we investigate a challenging unsupervised domain adaptation setting -- unsupervised model adaptation. We aim to explore how to rely only on unlabeled target data to improve performance of an existing source prediction model on the target domain, since labeled source data may not be available in some real-world scenarios due to data privacy issues. For this purpose, we propose a new framework, which is referred to as collaborative class conditional generative adversarial net to bypass the dependence on the source data. Specifically, the prediction model is to be improved through generated target-style data, which provides more accurate guidance for the generator. As a result, the generator and the prediction model can collaborate with each other without source data. Furthermore, due to the lack of supervision from source data, we propose a weight constraint that encourages similarity to the source model. A clustering-based regularization is also introduced to produce more discriminative features in the target domain. Compared to conventional domain adaptation methods, our model achieves superior performance on multiple adaptation tasks with only unlabeled target data, which verifies its effectiveness in this challenging setting.

accepted by CVPR2020

References in corpus (7)

• Conditional Generative Adversarial Nets
• How transferable are features in deep neural networks?
• Spectral Normalization for Generative Adversarial Networks
• Learning Transferable Features with Deep Adaptation Networks
• Deep Domain Confusion: Maximizing for Domain Invariance
• Domain Separation Networks
• A DIRT-T Approach to Unsupervised Domain Adaptation

Cited by in corpus (64)

• Tent: Fully Test-time Adaptation by Entropy Minimization
• A Comprehensive Survey on Test-Time Adaptation under Distribution Shifts
• Stagewise Unsupervised Domain Adaptation with Adversarial Self-Training for Road Segmentation of Remote Sensing Images
• Robust Multi-Task Learning and Online Refinement for Spacecraft Pose Estimation across Domain Gap
• Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation
• Exploiting the Intrinsic Neighborhood Structure for Source-free Domain Adaptation
• Model Adaptation: Historical Contrastive Learning for Unsupervised Domain Adaptation without Source Data
• FVP: Fourier Visual Prompting for Source-Free Unsupervised Domain Adaptation of Medical Image Segmentation
• Universal Domain Adaptation for Remote Sensing Image Scene Classification
• Uncertainty-Induced Transferability Representation for Source-Free Unsupervised Domain Adaptation
• KD3A: Unsupervised Multi-Source Decentralized Domain Adaptation via Knowledge Distillation
• Test-Time Adaptation to Distribution Shift by Confidence Maximization and Input Transformation
• Unsupervised Domain Adaptation of Black-Box Source Models
• Source-free Domain Adaptation via Distributional Alignment by Matching Batch Normalization Statistics
• A Prototype-Oriented Framework for Unsupervised Domain Adaptation
• Source Data-absent Unsupervised Domain Adaptation through Hypothesis Transfer and Labeling Transfer
• Uncertainty-Guided Mixup for Semi-Supervised Domain Adaptation without Source Data
• RAIN: RegulArization on Input and Network for Black-Box Domain Adaptation
• Transformer-Based Source-Free Domain Adaptation
• Exploiting Low-confidence Pseudo-labels for Source-free Object Detection
• Refining Language Models with Compositional Explanations
• Source-free Domain Adaptation via Avatar Prototype Generation and Adaptation
• Online Supervised Training of Spaceborne Vision during Proximity Operations using Adaptive Kalman Filtering
• Cross-domain Contrastive Learning for Unsupervised Domain Adaptation
• Source Free Unsupervised Graph Domain Adaptation
• Source-Free Domain Adaptation for Semantic Segmentation
• Attention Regularized Laplace Graph for Domain Adaptation
• Test time Adaptation through Perturbation Robustness
• A Free Lunch for Unsupervised Domain Adaptive Object Detection without Source Data
• Towards Unsupervised Domain Adaptation for Deep Face Recognition under Privacy Constraints via Federated Learning
• Generation, augmentation, and alignment: A pseudo-source domain based method for source-free domain adaptation
• VDM-DA: Virtual Domain Modeling for Source Data-free Domain Adaptation
• FixBi: Bridging Domain Spaces for Unsupervised Domain Adaptation
• Adaptive Domain Generalization via Online Disagreement Minimization
• Nearest Neighborhood-Based Deep Clustering for Source Data-absent Unsupervised Domain Adaptation
• DINE: Domain Adaptation from Single and Multiple Black-box Predictors
• Source-free Video Domain Adaptation by Learning from Noisy Labels
• Bridging the Domain Gap for Flight-Ready Spaceborne Vision
• Semi-Supervised Hypothesis Transfer for Source-Free Domain Adaptation
• Union-set Multi-source Model Adaptation for Semantic Segmentation
• Training on Test Data with Bayesian Adaptation for Covariate Shift
• Towards Corruption-Agnostic Robust Domain Adaptation
• Semi-Supervised Transfer Boosting (SS-TrBoosting)
• Source-Free Domain Adaptation for Question Answering with Masked Self-training
• Covariance-aware Feature Alignment with Pre-computed Source Statistics for Test-time Adaptation to Multiple Image Corruptions
• Consistent Assistant Domains Transformer for Source-free Domain Adaptation
• Cross-Domain Transfer Learning with CoRTe: Consistent and Reliable Transfer from Black-Box to Lightweight Segmentation Model
• Learning Invariant Representation with Consistency and Diversity for Semi-supervised Source Hypothesis Transfer
• Source-Free Adaptation to Measurement Shift via Bottom-Up Feature Restoration
• Visualizing Adapted Knowledge in Domain Transfer
• The Unreasonable Effectiveness of Large Language-Vision Models for Source-free Video Domain Adaptation
• Transfer Learning with Pre-trained Conditional Generative Models
• Domain Impression: A Source Data Free Domain Adaptation Method
• Uncover and Unlearn Nuisances: Agnostic Fully Test-Time Adaptation
• Domain Adaptation without Model Transferring
• Shuffle PatchMix Augmentation with Confidence-Margin Weighted Pseudo-Labels for Enhanced Source-Free Domain Adaptation
• Collaborative Semantic Aggregation and Calibration for Federated Domain Generalization
• Rethinking the Backbone in Class Imbalanced Federated Source Free Domain Adaptation: The Utility of Vision Foundation Models
• Taking Class Imbalance Into Account in Open Set Recognition Evaluation
• Adaptive Pseudo-Label Refinement by Negative Ensemble Learning for Source-Free Unsupervised Domain Adaptation
• Towards Data-Free Domain Generalization
• Knothe-Rosenblatt transport for Unsupervised Domain Adaptation
• Domain Adaptive Semantic Segmentation without Source Data
• On-target Adaptation