Co-regularized Alignment for Unsupervised Domain Adaptation

Abstract

Deep neural networks, trained with large amount of labeled data, can fail to generalize well when tested with examples from a \emph{target domain} whose distribution differs from the training data distribution, referred as the \emph{source domain}. It can be expensive or even infeasible to obtain required amount of labeled data in all possible domains. Unsupervised domain adaptation sets out to address this problem, aiming to learn a good predictive model for the target domain using labeled examples from the source domain but only unlabeled examples from the target domain. Domain alignment approaches this problem by matching the source and target feature distributions, and has been used as a key component in many state-of-the-art domain adaptation methods. However, matching the marginal feature distributions does not guarantee that the corresponding class conditional distributions will be aligned across the two domains. We propose co-regularized domain alignment for unsupervised domain adaptation, which constructs multiple diverse feature spaces and aligns source and target distributions in each of them individually, while encouraging that alignments agree with each other with regard to the class predictions on the unlabeled target examples. The proposed method is generic and can be used to improve any domain adaptation method which uses domain alignment. We instantiate it in the context of a recent state-of-the-art method and observe that it provides significant performance improvements on several domain adaptation benchmarks.

arXiv commentNIPS 2018 accepted version

Cited by in corpus (33)

• Deep Subdomain Adaptation Network for Image Classification
• Learning Robust Global Representations by Penalizing Local Predictive Power
• An Overview of Deep Semi-Supervised Learning
• Collaborative Unsupervised Domain Adaptation for Medical Image Diagnosis
• FDA: Fourier Domain Adaptation for Semantic Segmentation
• Virtual Mixup Training for Unsupervised Domain Adaptation
• Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation
• Transfer Learning with Dynamic Adversarial Adaptation Network
• Harmonizing Transferability and Discriminability for Adapting Object Detectors
• A Broader Study of Cross-Domain Few-Shot Learning
• Network Architecture Search for Domain Adaptation
• Label Propagation with Augmented Anchors: A Simple Semi-Supervised Learning baseline for Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation via Structured Prediction Based Selective Pseudo-Labeling
• Dual Adversarial Domain Adaptation
• Generation, augmentation, and alignment: A pseudo-source domain based method for source-free domain adaptation
• Accelerating Deep Unsupervised Domain Adaptation with Transfer Channel Pruning
• Contrastively Smoothed Class Alignment for Unsupervised Domain Adaptation
• Curriculum Manager for Source Selection in Multi-Source Domain Adaptation
• Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentation
• AdaFilter: Adaptive Filter Fine-tuning for Deep Transfer Learning
• A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation
• Unsupervised Domain Adaptation with Progressive Domain Augmentation
• Phase Consistent Ecological Domain Adaptation
• Joint Semantic Domain Alignment and Target Classifier Learning for Unsupervised Domain Adaptation
• Hypothesis Disparity Regularized Mutual Information Maximization
• Learning Classifiers for Domain Adaptation, Zero and Few-Shot Recognition Based on Learning Latent Semantic Parts
• Enlarging Discriminative Power by Adding an Extra Class in Unsupervised Domain Adaptation
• Mixup Regularized Adversarial Networks for Multi-Domain Text Classification
• A Dictionary Approach to Domain-Invariant Learning in Deep Networks
• Multi-Purposing Domain Adaptation Discriminators for Pseudo Labeling Confidence
• ADeLA: Automatic Dense Labeling with Attention for Viewpoint Adaptation in Semantic Segmentation
• Unsupervised Domain Adaptive Object Detection using Forward-Backward Cyclic Adaptation
• iFAN: Image-Instance Full Alignment Networks for Adaptive Object Detection