Unsupervised Domain Adaptation with Residual Transfer Networks

Abstract

The recent success of deep neural networks relies on massive amounts of labeled data. For a target task where labeled data is unavailable, domain adaptation can transfer a learner from a different source domain. In this paper, we propose a new approach to domain adaptation in deep networks that can jointly learn adaptive classifiers and transferable features from labeled data in the source domain and unlabeled data in the target domain. We relax a shared-classifier assumption made by previous methods and assume that the source classifier and target classifier differ by a residual function. We enable classifier adaptation by plugging several layers into deep network to explicitly learn the residual function with reference to the target classifier. We fuse features of multiple layers with tensor product and embed them into reproducing kernel Hilbert spaces to match distributions for feature adaptation. The adaptation can be achieved in most feed-forward models by extending them with new residual layers and loss functions, which can be trained efficiently via back-propagation. Empirical evidence shows that the new approach outperforms state of the art methods on standard domain adaptation benchmarks.

arXiv comment30th Conference on Neural Information Processing Systems (NIPS 2016), Barcelona, Spain

References in corpus (7)

• Natural Language Processing (almost) from Scratch
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• How transferable are features in deep neural networks?
• Learning Transferable Features with Deep Adaptation Networks
• Deep Domain Confusion: Maximizing for Domain Invariance
• Domain Adaptation: Learning Bounds and Algorithms
• LSDA: Large Scale Detection Through Adaptation

Cited by in corpus (302)

• Deep Visual Domain Adaptation: A Survey
• ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
• A review of domain adaptation without target labels
• Deep Reinforcement Learning: An Overview
• Model Adaptation: Unsupervised Domain Adaptation without Source Data
• Beyond Sharing Weights for Deep Domain Adaptation
• Asymmetric Tri-training for Unsupervised Domain Adaptation
• Kernel Mean Embedding of Distributions: A Review and Beyond
• Learning Decentralized Controllers for Robot Swarms with Graph Neural Networks
• Aligning Domain-specific Distribution and Classifier for Cross-domain Classification from Multiple Sources
• Revisiting Batch Normalization For Practical Domain Adaptation
• A Survey on Incorporating Domain Knowledge into Deep Learning for Medical Image Analysis
• Learning Robust Global Representations by Penalizing Local Predictive Power
• A Comprehensive Survey on Transfer Learning
• What, Where and How to Transfer in SAR Target Recognition Based on Deep CNNs
• Collaborative Unsupervised Domain Adaptation for Medical Image Diagnosis
• Deep Residual Correction Network for Partial Domain Adaptation
• Self-supervised Domain Adaptation for Computer Vision Tasks
• On Learning Invariant Representation for Domain Adaptation
• Heterogeneous domain adaptation: An unsupervised approach
• Multiscale Domain Adaptive YOLO for Cross-Domain Object Detection
• Automatic Sleep Staging of EEG Signals: Recent Development, Challenges, and Future Directions
• Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Learning Disentangled Semantic Representation for Domain Adaptation
• Learning to cluster in order to transfer across domains and tasks
• Deep Transfer Learning with Joint Adaptation Networks
• Enabling Country-Scale Land Cover Mapping with Meter-Resolution Satellite Imagery
• Unsupervised Domain Adaptation with Adversarial Residual Transform Networks
• Unsupervised Multi-Class Domain Adaptation: Theory, Algorithms, and Practice
• Contrastive Adaptation Network for Unsupervised Domain Adaptation
• Robust Unsupervised Domain Adaptation for Neural Networks via Moment Alignment
• Model Adaptation: Historical Contrastive Learning for Unsupervised Domain Adaptation without Source Data
• Cross-regional oil palm tree counting and detection via multi-level attention domain adaptation network
• Multi-Adversarial Domain Adaptation
• Demystifying Neural Style Transfer
• Optimal Bayesian Transfer Learning
• SCL: Towards Accurate Domain Adaptive Object Detection via Gradient Detach Based Stacked Complementary Losses
• Multi-label Learning Based Deep Transfer Neural Network for Facial Attribute Classification
• Feature-Critic Networks for Heterogeneous Domain Generalization
• Test-Time Training with Self-Supervision for Generalization under Distribution Shifts
• Generalized Domain Conditioned Adaptation Network
• Pseudo-Labeling Curriculum for Unsupervised Domain Adaptation
• Domain Adaptation with Conditional Distribution Matching and Generalized Label Shift
• Transfer Learning via Contextual Invariants for One-to-Many Cross-Domain Recommendation
• Self-Supervised Policy Adaptation during Deployment
• Cycle Self-Training for Domain Adaptation
• Learning Transferable Parameters for Unsupervised Domain Adaptation
• Average Biased ReLU Based CNN Descriptor for Improved Face Retrieval
• Small Sample Learning in Big Data Era
• Efficient Convolutional Neural Networks for Depth-Based Multi-Person Pose Estimation
• Dynamic Instance Domain Adaptation
• Frustratingly Simple Domain Generalization via Image Stylization
• Deep Cocktail Network: Multi-source Unsupervised Domain Adaptation with Category Shift
• Adversarial Dropout Regularization
• Transferrable Prototypical Networks for Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation for Mobile Semantic Segmentation based on Cycle Consistency and Feature Alignment
• Moment Matching for Multi-Source Domain Adaptation
• Larger Norm More Transferable: An Adaptive Feature Norm Approach for Unsupervised Domain Adaptation
• Learning Target Domain Specific Classifier for Partial Domain Adaptation
• Multi-adversarial Faster-RCNN for Unrestricted Object Detection
• Domain-Symmetric Networks for Adversarial Domain Adaptation
• A Survey on Deep Transfer Learning
• Unsupervised Domain Adaptation for Object Detection via Cross-Domain Semi-Supervised Learning
• Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptation
• Inferring Latent Domains for Unsupervised Deep Domain Adaptation
• Multi-target Unsupervised Domain Adaptation without Exactly Shared Categories
• Learning to Transfer Examples for Partial Domain Adaptation
• Feature Alignment and Restoration for Domain Generalization and Adaptation
• Cross-domain Self-supervised Learning for Domain Adaptation with Few Source Labels
• Confidence Regularized Self-Training
• Racial Faces in-the-Wild: Reducing Racial Bias by Information Maximization Adaptation Network
• Domain Adaptive Attention Learning for Unsupervised Person Re-Identification
• Transfer Learning with Dynamic Adversarial Adaptation Network
• Drop to Adapt: Learning Discriminative Features for Unsupervised Domain Adaptation
• Exploring Object Relation in Mean Teacher for Cross-Domain Detection
• An Unsupervised Domain Adaptation Model based on Dual-module Adversarial Training
• Constrained domain adaptation for Image segmentation
• Partial Adversarial Domain Adaptation
• FSDR: Frequency Space Domain Randomization for Domain Generalization
• Episodic Training for Domain Generalization
• Domain Aggregation Networks for Multi-Source Domain Adaptation
• Domain-Specific Batch Normalization for Unsupervised Domain Adaptation
• Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations
• Minimum Class Confusion for Versatile Domain Adaptation
• Domain Adaptation without Source Data
• A Prototype-Oriented Framework for Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation via Structurally Regularized Deep Clustering
• Birds of A Feather Flock Together: Category-Divergence Guidance for Domain Adaptive Segmentation
• Progressive Feature Alignment for Unsupervised Domain Adaptation
• No More Discrimination: Cross City Adaptation of Road Scene Segmenters
• Open Set Domain Adaptation by Backpropagation
• Cluster Alignment with a Teacher for Unsupervised Domain Adaptation
• Attract, Perturb, and Explore: Learning a Feature Alignment Network for Semi-supervised Domain Adaptation
• Semantic Distribution-aware Contrastive Adaptation for Semantic Segmentation
• Transformer-Based Source-Free Domain Adaptation
• Importance Weighted Adversarial Nets for Partial Domain Adaptation
• Rethink Maximum Mean Discrepancy for Domain Adaptation
• Unbiased Mean Teacher for Cross-domain Object Detection
• Coupling Adversarial Learning with Selective Voting Strategy for Distribution Alignment in Partial Domain Adaptation
• Transferable Calibration with Lower Bias and Variance in Domain Adaptation
• Cross-Resolution Adversarial Dual Network for Person Re-Identification and Beyond
• Regularizing Proxies with Multi-Adversarial Training for Unsupervised Domain-Adaptive Semantic Segmentation
• A Novel Unsupervised Camera-aware Domain Adaptation Framework for Person Re-identification
• Temporal Attentive Alignment for Large-Scale Video Domain Adaptation
• Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models
• Zero-shot Domain Adaptation without Domain Semantic Descriptors
• Induction Networks for Few-Shot Text Classification
• TWINs: Two Weighted Inconsistency-reduced Networks for Partial Domain Adaptation
• Domain Adaptive Object Detection via Asymmetric Tri-way Faster-RCNN
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• Noise Adaptive Speech Enhancement using Domain Adversarial Training
• Semi-supervised Models are Strong Unsupervised Domain Adaptation Learners
• Domain-Invariant Adversarial Learning for Unsupervised Domain Adaption
• Bringing Giant Neural Networks Down to Earth with Unlabeled Data
• Minimax Lower Bounds for Transfer Learning with Linear and One-hidden Layer Neural Networks
• Universal Semi-Supervised Semantic Segmentation
• Modified Distribution Alignment for Domain Adaptation with Pre-trained Inception ResNet
• Entropy Minimization vs. Diversity Maximization for Domain Adaptation
• Open-Set Hypothesis Transfer with Semantic Consistency
• Mutual Learning Network for Multi-Source Domain Adaptation
• Estimating Generalization under Distribution Shifts via Domain-Invariant Representations
• Unifying Unsupervised Domain Adaptation and Zero-Shot Visual Recognition
• Learning Factorized Representations for Open-set Domain Adaptation
• Dual-Refinement: Joint Label and Feature Refinement for Unsupervised Domain Adaptive Person Re-Identification
• CLDA: Contrastive Learning for Semi-Supervised Domain Adaptation
• Action Segmentation with Joint Self-Supervised Temporal Domain Adaptation
• When Low Resource NLP Meets Unsupervised Language Model: Meta-pretraining Then Meta-learning for Few-shot Text Classification
• Adversarial Alignment of Class Prediction Uncertainties for Domain Adaptation
• Contextual-Relation Consistent Domain Adaptation for Semantic Segmentation
• SPG: Unsupervised Domain Adaptation for 3D Object Detection via Semantic Point Generation
• Multi-Source Domain Adaptation with Collaborative Learning for Semantic Segmentation
• Few shot domain adaptation for in situ macromolecule structural classification in cryo-electron tomograms
• Label Propagation with Augmented Anchors: A Simple Semi-Supervised Learning baseline for Unsupervised Domain Adaptation
• Improving Compound Activity Classification via Deep Transfer and Representation Learning
• Category Contrast for Unsupervised Domain Adaptation in Visual Tasks
• Unsupervised Domain Adaptation via Structured Prediction Based Selective Pseudo-Labeling
• Improving Unsupervised Domain Adaptation with Variational Information Bottleneck
• DART: Domain-Adversarial Residual-Transfer Networks for Unsupervised Cross-Domain Image Classification
• Dual Adversarial Domain Adaptation
• Foreground-Aware Stylization and Consensus Pseudo-Labeling for Domain Adaptation of First-Person Hand Segmentation
• Semantic-Transferable Weakly-Supervised Endoscopic Lesions Segmentation
• On the Exploration of Incremental Learning for Fine-grained Image Retrieval
• Learning to Combine: Knowledge Aggregation for Multi-Source Domain Adaptation
• MiniMax Entropy Network: Learning Category-Invariant Features for Domain Adaptation
• Associative Partial Domain Adaptation
• Domain Adaptive Medical Image Segmentation via Adversarial Learning of Disease-Specific Spatial Patterns
• Adversarial Camera Alignment Network for Unsupervised Cross-camera Person Re-identification
• Domain Adaptation with Auxiliary Target Domain-Oriented Classifier
• Domain Attention Consistency for Multi-Source Domain Adaptation
• Known-class Aware Self-ensemble for Open Set Domain Adaptation
• Cross-Domain Gradient Discrepancy Minimization for Unsupervised Domain Adaptation
• AdapNet: Adaptability Decomposing Encoder-Decoder Network for Weakly Supervised Action Recognition and Localization
• Adversarial Cross-Domain Action Recognition with Co-Attention
• Attending to Discriminative Certainty for Domain Adaptation
• Collaborative Training of Balanced Random Forests for Open Set Domain Adaptation
• Unsupervised Multi-Source Domain Adaptation for Person Re-Identification
• Contrastively Smoothed Class Alignment for Unsupervised Domain Adaptation
• Style Normalization and Restitution for Domain Generalization and Adaptation
• Cross-Domain Adaptation for Animal Pose Estimation
• Data Augmentation with norm-VAE for Unsupervised Domain Adaptation
• Deep Adversarial Transition Learning using Cross-Grafted Generative Stacks
• Multi-source Distilling Domain Adaptation
• Curriculum based Dropout Discriminator for Domain Adaptation
• Disjoint Label Space Transfer Learning with Common Factorised Space
• Be Like Water: Robustness to Extraneous Variables Via Adaptive Feature Normalization
• A Sample Selection Approach for Universal Domain Adaptation
• Normalized Wasserstein Distance for Mixture Distributions with Applications in Adversarial Learning and Domain Adaptation
• Hand-Object Contact Consistency Reasoning for Human Grasps Generation
• Dynamic Scale Inference by Entropy Minimization
• Domain-Adaptive Few-Shot Learning
• Learning Smooth Representation for Unsupervised Domain Adaptation
• The Battleship Approach to the Low Resource Entity Matching Problem
• Learning to Match Distributions for Domain Adaptation
• Online Meta-Learning for Multi-Source and Semi-Supervised Domain Adaptation
• Adversarial Domain Adaptation for Machine Reading Comprehension
• MetaAlign: Coordinating Domain Alignment and Classification for Unsupervised Domain Adaptation
• Stretching Domain Adaptation: How far is too far?
• Collaborative Training between Region Proposal Localization and Classification for Domain Adaptive Object Detection
• Discriminative Cross-Domain Feature Learning for Partial Domain Adaptation
• Bi-Classifier Determinacy Maximization for Unsupervised Domain Adaptation
• Understanding the Limits of Unsupervised Domain Adaptation via Data Poisoning
• Semi-Supervised Domain Adaptation via Adaptive and Progressive Feature Alignment
• Selective Transfer with Reinforced Transfer Network for Partial Domain Adaptation
• Discriminative Adversarial Domain Adaptation
• SHAPED: Shared-Private Encoder-Decoder for Text Style Adaptation
• Secure Domain Adaptation with Multiple Sources
• Domain-Specific Suppression for Adaptive Object Detection
• A Cross-Level Information Transmission Network for Predicting Phenotype from New Genotype: Application to Cancer Precision Medicine
• Weakly Supervised Open-set Domain Adaptation by Dual-domain Collaboration
• Mind the Gap: Enlarging the Domain Gap in Open Set Domain Adaptation
• Pixel and Feature Level Based Domain Adaption for Object Detection in Autonomous Driving
• AutoFT: Automatic Fine-Tune for Parameters Transfer Learning in Click-Through Rate Prediction
• Domain Adaptation with Randomized Expectation Maximization
• Cycle-consistent Conditional Adversarial Transfer Networks
• Domain Generalization via Inference-time Label-Preserving Target Projections
• Joint Semantic Domain Alignment and Target Classifier Learning for Unsupervised Domain Adaptation
• MADAN: Multi-source Adversarial Domain Aggregation Network for Domain Adaptation
• STEP: Spatio-Temporal Progressive Learning for Video Action Detection
• Domain Adaptation and Image Classification via Deep Conditional Adaptation Network
• Moment-Based Domain Adaptation: Learning Bounds and Algorithms
• Deeply Aligned Adaptation for Cross-domain Object Detection
• Multiple Classifiers Based Maximum Classifier Discrepancy for Unsupervised Domain Adaptation
• Multi-Source Domain Adaptation for Text Classification via DistanceNet-Bandits
• GradMix: Multi-source Transfer across Domains and Tasks
• Adaptively-Accumulated Knowledge Transfer for Partial Domain Adaptation
• ADMP: An Adversarial Double Masks Based Pruning Framework For Unsupervised Cross-Domain Compression
• Unsupervised Domain Adaptation with Temporal-Consistent Self-Training for 3D Hand-Object Joint Reconstruction
• Sparse-to-dense Feature Matching: Intra and Inter domain Cross-modal Learning in Domain Adaptation for 3D Semantic Segmentation
• AutoAdapt: Automated Segmentation Network Search for Unsupervised Domain Adaptation
• Domain Adaptive SiamRPN++ for Object Tracking in the Wild
• Next-Gen Machine Learning Supported Diagnostic Systems for Spacecraft
• Preserving Semantic Consistency in Unsupervised Domain Adaptation Using Generative Adversarial Networks
• S2R-DepthNet: Learning a Generalizable Depth-specific Structural Representation
• Using Latent Codes for Class Imbalance Problem in Unsupervised Domain Adaptation
• Deep Cross-media Knowledge Transfer
• Dynamic Filtering with Large Sampling Field for ConvNets
• Gotta Adapt 'Em All: Joint Pixel and Feature-Level Domain Adaptation for Recognition in the Wild
• Domain Adaptation Meets Disentangled Representation Learning and Style Transfer
• A Compact DNN: Approaching GoogLeNet-Level Accuracy of Classification and Domain Adaptation
• Self-Supervised Domain Adaptation with Consistency Training
• Subdomain Adaptation with Manifolds Discrepancy Alignment
• Importance Weighted Adversarial Discriminative Transfer for Anomaly Detection
• Conditional Coupled Generative Adversarial Networks for Zero-Shot Domain Adaptation
• Sequential Learning for Domain Generalization
• Multimodal Deep Domain Adaptation
• Stochastic Adversarial Gradient Embedding for Active Domain Adaptation
• Multi-source Few-shot Domain Adaptation
• Multilevel Knowledge Transfer for Cross-Domain Object Detection
• Uncovering the Connections Between Adversarial Transferability and Knowledge Transferability
• Spectral Unsupervised Domain Adaptation for Visual Recognition
• Discriminative Clustering for Robust Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation using Deep Networks with Cross-Grafted Stacks
• Deep Learning for HDR Imaging: State-of-the-Art and Future Trends
• Impact of ImageNet Model Selection on Domain Adaptation
• Target Consistency for Domain Adaptation: when Robustness meets Transferability
• Domain Contrast for Domain Adaptive Object Detection
• Attract or Distract: Exploit the Margin of Open Set
• Unveiling Class-Labeling Structure for Universal Domain Adaptation
• Teacher-Student Consistency For Multi-Source Domain Adaptation
• Efficient Pre-trained Features and Recurrent Pseudo-Labeling in Unsupervised Domain Adaptation
• More Separable and Easier to Segment: A Cluster Alignment Method for Cross-Domain Semantic Segmentation
• Joint Distribution Alignment via Adversarial Learning for Domain Adaptive Object Detection
• Collaborative Semantic Aggregation and Calibration for Federated Domain Generalization
• Semantic-Rearrangement-Based Multi-Level Alignment for Domain Generalized Segmentation
• On Direct Distribution Matching for Adapting Segmentation Networks
• Self-Adaptive Partial Domain Adaptation
• Adaptive Distillation: Aggregating Knowledge from Multiple Paths for Efficient Distillation
• PCA-based Multi Task Learning: a Random Matrix Approach
• Against Adversarial Learning: Naturally Distinguish Known and Unknown in Open Set Domain Adaptation
• Interventional Domain Adaptation
• Sim2Real for Self-Supervised Monocular Depth and Segmentation
• Domain Adaptation on Semantic Segmentation for Aerial Images
• Towards Recognizing New Semantic Concepts in New Visual Domains
• Exploiting Images for Video Recognition with Hierarchical Generative Adversarial Networks
• Domain2Vec: Domain Embedding for Unsupervised Domain Adaptation
• Unsupervised Domain Adaptation: A Multi-task Learning-based Method
• Manifold Criterion Guided Transfer Learning via Intermediate Domain Generation
• Discrepancy Minimization in Domain Generalization with Generative Nearest Neighbors
• Adversarial Dual Distinct Classifiers for Unsupervised Domain Adaptation
• Adversarial Learning for Zero-shot Domain Adaptation
• Range Adaptation for 3D Object Detection in LiDAR
• Contrastive Language Adaptation for Cross-Lingual Stance Detection
• A Dictionary Approach to Domain-Invariant Learning in Deep Networks
• Cross-domain error minimization for unsupervised domain adaptation
• Dynamic Adaptation on Non-Stationary Visual Domains
• Enhanced Separable Disentanglement for Unsupervised Domain Adaptation
• Ensemble Multi-Source Domain Adaptation with Pseudolabels
• Importance Filtered Cross-Domain Adaptation
• SpotPatch: Parameter-Efficient Transfer Learning for Mobile Object Detection
• Self-Guided Adaptation: Progressive Representation Alignment for Domain Adaptive Object Detection
• Transferable Semantic Augmentation for Domain Adaptation
• Geometry-Aware Unsupervised Domain Adaptation for Stereo Matching
• Generalized Domain Adaptation
• Correlated Adversarial Joint Discrepancy Adaptation Network
• Unsupervised Domain Adaptation: A Reality Check
• Covariate Shift in High-Dimensional Random Feature Regression
• Domain Adaptation with Incomplete Target Domains
• The Role of Embedding Complexity in Domain-invariant Representations
• DWMD: Dimensional Weighted Orderwise Moment Discrepancy for Domain-specific Hidden Representation Matching
• A Survey on Green Deep Learning
• Learning Task-oriented Disentangled Representations for Unsupervised Domain Adaptation
• Learning to see across Domains and Modalities
• CrDoCo: Pixel-level Domain Transfer with Cross-Domain Consistency
• Unsupervised Domain Adaptive Object Detection using Forward-Backward Cyclic Adaptation
• Exploring Dropout Discriminator for Domain Adaptation
• Global and Local Texture Randomization for Synthetic-to-Real Semantic Segmentation
• Exploring Data Aggregation and Transformations to Generalize across Visual Domains
• Classifying In-Place Gestures with End-to-End Point Cloud Learning
• Densely Semantic Enhancement for Domain Adaptive Region-free Detectors
• Domain Adaptation by Maximizing Population Correlation with Neural Architecture Search
• Dataset Bias in Few-shot Image Recognition
• 3D Scene Parsing via Class-Wise Adaptation
• Adapting Vehicle Detector to Target Domain by Adversarial Prediction Alignment
• Efficient Video Understanding via Layered Multi Frame-Rate Analysis
• Deep Spherical Manifold Gaussian Kernel for Unsupervised Domain Adaptation
• Open Set Domain Adaptation for Image and Action Recognition
• Deep Least Squares Alignment for Unsupervised Domain Adaptation
• Cross Domain Image Matching in Presence of Outliers
• Universal Multi-Source Domain Adaptation
• Adversarial Transfer Learning for Cross-domain Visual Recognition
• A Novel Transferability Attention Neural Network Model for EEG Emotion Recognition
• Towards Uncovering the Intrinsic Data Structures for Unsupervised Domain Adaptation using Structurally Regularized Deep Clustering