Deep CORAL: Correlation Alignment for Deep Domain Adaptation

Abstract

Deep neural networks are able to learn powerful representations from large quantities of labeled input data, however they cannot always generalize well across changes in input distributions. Domain adaptation algorithms have been proposed to compensate for the degradation in performance due to domain shift. In this paper, we address the case when the target domain is unlabeled, requiring unsupervised adaptation. CORAL is a "frustratingly easy" unsupervised domain adaptation method that aligns the second-order statistics of the source and target distributions with a linear transformation. Here, we extend CORAL to learn a nonlinear transformation that aligns correlations of layer activations in deep neural networks (Deep CORAL). Experiments on standard benchmark datasets show state-of-the-art performance.

Extended Abstract

References in corpus (4)

• Caffe: Convolutional Architecture for Fast Feature Embedding
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• Deep Domain Confusion: Maximizing for Domain Invariance
• Learning from Multiple Outlooks

Cited by in corpus (69)

• Deep Visual Domain Adaptation: A Survey
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Adversarial Discriminative Domain Adaptation
• Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sources
• Revisiting Batch Normalization For Practical Domain Adaptation
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• Multi-Representation Adaptation Network for Cross-domain Image Classification
• Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learning
• Few-Shot Adversarial Domain Adaptation
• Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Deep Domain Adaptation Based Video Smoke Detection using Synthetic Smoke Images
• In Search of Lost Domain Generalization
• Label Efficient Learning of Transferable Representations across Domains and Tasks
• Adversarial Dropout Regularization
• Mid-Level Visual Representations Improve Generalization and Sample Efficiency for Learning Visuomotor Policies
• Associative Domain Adaptation
• AdaRNN: Adaptive Learning and Forecasting of Time Series
• Unsupervised Cross-Modality Domain Adaptation of ConvNets for Biomedical Image Segmentations with Adversarial Loss
• Joint Domain Alignment and Discriminative Feature Learning for Unsupervised Deep Domain Adaptation
• Domain Adaptation as a Problem of Inference on Graphical Models
• Domain Adaptation Principal Component Analysis: base linear method for learning with out-of-distribution data
• Causal Generative Domain Adaptation Networks
• Importance Weighted Adversarial Nets for Partial Domain Adaptation
• Joint Disentangling and Adaptation for Cross-Domain Person Re-Identification
• All about Structure: Adapting Structural Information across Domains for Boosting Semantic Segmentation
• Characterizing and Avoiding Negative Transfer
• Prototypical Partial Optimal Transport for Universal Domain Adaptation
• Adaptive Semantic Segmentation with a Strategic Curriculum of Proxy Labels
• Correlation Alignment by Riemannian Metric for Domain Adaptation
• Synthetic to Real Adaptation with Generative Correlation Alignment Networks
• Domain-Invariant Adversarial Learning for Unsupervised Domain Adaption
• POEM: Polarization of Embeddings for Domain-Invariant Representations
• ACE: Adapting to Changing Environments for Semantic Segmentation
• Coupling weak and strong supervision for classification of prostate cancer histopathology images
• Learning Factorized Representations for Open-set Domain Adaptation
• Domain Alignment with Triplets
• Disentangled Variational Representation for Heterogeneous Face Recognition
• Domain Adaptation from Synthesis to Reality in Single-model Detector for Video Smoke Detection
• Associative Partial Domain Adaptation
• Deep causal representation learning for unsupervised domain adaptation
• Adversarial Attribute-Image Person Re-identification
• Deep Domain Adaptation by Geodesic Distance Minimization
• Exploiting Local Feature Patterns for Unsupervised Domain Adaptation
• Factorized Adversarial Networks for Unsupervised Domain Adaptation
• GA-DAN: Geometry-Aware Domain Adaptation Network for Scene Text Detection and Recognition
• Cycle-consistent Conditional Adversarial Transfer Networks
• Structure-Attentioned Memory Network for Monocular Depth Estimation
• Domain Adaptation Meets Disentangled Representation Learning and Style Transfer
• Domain Conditioned Adaptation Network
• Dynamic Feature Alignment for Semi-supervised Domain Adaptation
• Contradictory Structure Learning for Semi-supervised Domain Adaptation
• Unsupervised Domain Adaptation using Generative Models and Self-ensembling
• Uncorrelated Feature Encoding for Faster Image Style Transfer
• Visual Domain Adaptation with Manifold Embedded Distribution Alignment
• Spatial-Aware GAN for Unsupervised Person Re-identification
• Uniform Priors for Data-Efficient Transfer
• Deep unsupervised domain adaptation applied to the Cherenkov Telescope Array Large-Sized Telescope
• Coverage-Guaranteed Prediction Sets for Out-of-Distribution Data
• A Fully Convolutional Tri-branch Network (FCTN) for Domain Adaptation
• Exploiting Images for Video Recognition with Hierarchical Generative Adversarial Networks
• Deep Discriminative Learning for Unsupervised Domain Adaptation
• Regularized Wasserstein Means for Aligning Distributional Data
• Transfer Learning using Representation Learning in Massive Open Online Courses
• Unbiased Image Style Transfer
• Parameter Reference Loss for Unsupervised Domain Adaptation
• Dynamic Adaptation on Non-Stationary Visual Domains
• Multi-component Image Translation for Deep Domain Generalization
• Task Guided Compositional Representation Learning for ZDA
• KL Guided Domain Adaptation