Wasserstein Distance Guided Representation Learning for Domain Adaptation

Abstract

Domain adaptation aims at generalizing a high-performance learner on a target domain via utilizing the knowledge distilled from a source domain which has a different but related data distribution. One solution to domain adaptation is to learn domain invariant feature representations while the learned representations should also be discriminative in prediction. To learn such representations, domain adaptation frameworks usually include a domain invariant representation learning approach to measure and reduce the domain discrepancy, as well as a discriminator for classification. Inspired by Wasserstein GAN, in this paper we propose a novel approach to learn domain invariant feature representations, namely Wasserstein Distance Guided Representation Learning (WDGRL). WDGRL utilizes a neural network, denoted by the domain critic, to estimate empirical Wasserstein distance between the source and target samples and optimizes the feature extractor network to minimize the estimated Wasserstein distance in an adversarial manner. The theoretical advantages of Wasserstein distance for domain adaptation lie in its gradient property and promising generalization bound. Empirical studies on common sentiment and image classification adaptation datasets demonstrate that our proposed WDGRL outperforms the state-of-the-art domain invariant representation learning approaches.

The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI 2018)

Cited by in corpus (122)

• A review of domain adaptation without target labels
• DeepJDOT: Deep Joint Distribution Optimal Transport for Unsupervised Domain Adaptation
• WILDS: A Benchmark of in-the-Wild Distribution Shifts
• Wasserstein Distance based Deep Adversarial Transfer Learning for Intelligent Fault Diagnosis
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• A Comprehensive Survey on Transfer Learning
• On Learning Invariant Representation for Domain Adaptation
• Heterogeneous domain adaptation: An unsupervised approach
• Self-supervised Autoregressive Domain Adaptation for Time Series Data
• Towards Accurate Model Selection in Deep Unsupervised Domain Adaptation
• Domain Generalization via Optimal Transport with Metric Similarity Learning
• Domain Adaptation with Conditional Distribution Matching and Generalized Label Shift
• Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment
• Bridging the Gap between Events and Frames through Unsupervised Domain Adaptation
• Robust Optimal Transport with Applications in Generative Modeling and Domain Adaptation
• Rethinking Distributional Matching Based Domain Adaptation
• Adversarial Training in Affective Computing and Sentiment Analysis: Recent Advances and Perspectives
• Unbalanced minibatch Optimal Transport; applications to Domain Adaptation
• Cross-domain Self-supervised Learning for Domain Adaptation with Few Source Labels
• CoDA: Contrast-enhanced and Diversity-promoting Data Augmentation for Natural Language Understanding
• Learning to Map Natural Language Instructions to Physical Quadcopter Control using Simulated Flight
• Domain-Specific Batch Normalization for Unsupervised Domain Adaptation
• Unsupervised Neural Domain Adaptation for Document Image Binarization
• Harmonizing Transferability and Discriminability for Adapting Object Detectors
• Semi-supervised representation learning via dual autoencoders for domain adaptation
• Robust Federated Learning Through Representation Matching and Adaptive Hyper-parameters
• See Eye to Eye: A Lidar-Agnostic 3D Detection Framework for Unsupervised Multi-Target Domain Adaptation
• A survey on domain adaptation theory: learning bounds and theoretical guarantees
• Towards Fair Knowledge Transfer for Imbalanced Domain Adaptation
• Cross-domain Detection via Graph-induced Prototype Alignment
• Empirically Measuring Transfer Distance for System Design and Operation
• Bridging the Theoretical Bound and Deep Algorithms for Open Set Domain Adaptation
• Corner Cases for Visual Perception in Automated Driving: Some Guidance on Detection Approaches
• Channel adversarial training for speaker verification and diarization
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• Minimax Lower Bounds for Transfer Learning with Linear and One-hidden Layer Neural Networks
• CLDA: Contrastive Learning for Semi-Supervised Domain Adaptation
• Estimating Generalization under Distribution Shifts via Domain-Invariant Representations
• Label Propagation with Augmented Anchors: A Simple Semi-Supervised Learning baseline for Unsupervised Domain Adaptation
• SharinGAN: Combining Synthetic and Real Data for Unsupervised Geometry Estimation
• Online Adaptation to Label Distribution Shift
• Do Neural Optimal Transport Solvers Work? A Continuous Wasserstein-2 Benchmark
• A Survey on Deep Domain Adaptation and Tiny Object Detection Challenges, Techniques and Datasets
• VDM-DA: Virtual Domain Modeling for Source Data-free Domain Adaptation
• Generation, augmentation, and alignment: A pseudo-source domain based method for source-free domain adaptation
• Open Set Domain Adaptation: Theoretical Bound and Algorithm
• Attending to Discriminative Certainty for Domain Adaptation
• Learning Domain Invariant Representations by Joint Wasserstein Distance Minimization
• Semantics-Aware Image to Image Translation and Domain Transfer
• On Universal Black-Box Domain Adaptation
• Concentration bounds for linear Monge mapping estimation and optimal transport domain adaptation
• Transfer Learning for Relation Extraction via Relation-Gated Adversarial Learning
• An Adversarial Domain Separation Framework for Septic Shock Early Prediction Across EHR Systems
• Curriculum based Dropout Discriminator for Domain Adaptation
• Class-imbalanced Domain Adaptation: An Empirical Odyssey
• Model-based Policy Optimization with Unsupervised Model Adaptation
• Continuous Transfer Learning with Label-informed Distribution Alignment
• DAN: Dual-View Representation Learning for Adapting Stance Classifiers to New Domains
• Adaptive Object Detection with Dual Multi-Label Prediction
• Domain-Specific Suppression for Adaptive Object Detection
• Unsupervised Domain Adaptation for Image Classification via Structure-Conditioned Adversarial Learning
• Understanding the Limits of Unsupervised Domain Adaptation via Data Poisoning
• From Shallow to Deep Interactions Between Knowledge Representation, Reasoning and Machine Learning (Kay R. Amel group)
• Exploiting Local Feature Patterns for Unsupervised Domain Adaptation
• A Cross-Level Information Transmission Network for Predicting Phenotype from New Genotype: Application to Cancer Precision Medicine
• Augmented Sliced Wasserstein Distances
• Domain Adversarial Reinforcement Learning for Partial Domain Adaptation
• Wasserstein-Fisher-Rao Document Distance
• MADAN: Multi-source Adversarial Domain Aggregation Network for Domain Adaptation
• Towards Discriminative Representation Learning for Unsupervised Person Re-identification
• G2DA: Geometry-Guided Dual-Alignment Learning for RGB-Infrared Person Re-Identification
• Privacy Enhancing Machine Learning via Removal of Unwanted Dependencies
• Multi-Source Domain Adaptation with Mixture of Experts
• Multi-Source Deep Domain Adaptation for Quality Control in Retail Food Packaging
• Improving Unsupervised Domain Adaptation by Reducing Bi-level Feature Redundancy
• Domain Adaptation for Sentiment Analysis Using Increased Intraclass Separation
• Mitigating Uncertainty of Classifier for Unsupervised Domain Adaptation
• CNN-based Discriminative Training for Domain Compensation in Acoustic Event Detection with Frame-wise Classifier
• Improving Domain-Adapted Sentiment Classification by Deep Adversarial Mutual Learning
• Hidden Covariate Shift: A Minimal Assumption For Domain Adaptation
• Relation Adversarial Network for Low Resource Knowledge Graph Completion
• Multi-Source Domain Adaptation for Text Classification via DistanceNet-Bandits
• Meta Fine-Tuning Neural Language Models for Multi-Domain Text Mining
• Adversarial Deep Network Embedding for Cross-network Node Classification
• Representation Bayesian Risk Decompositions and Multi-Source Domain Adaptation
• EasyTransfer -- A Simple and Scalable Deep Transfer Learning Platform for NLP Applications
• Multi-source Few-shot Domain Adaptation
• Curriculum CycleGAN for Textual Sentiment Domain Adaptation with Multiple Sources
• Domain Impression: A Source Data Free Domain Adaptation Method
• Multi-task Learning by Leveraging the Semantic Information
• Discriminative Active Learning for Domain Adaptation
• Adv-4-Adv: Thwarting Changing Adversarial Perturbations via Adversarial Domain Adaptation
• Synthetic training data generation for deep learning based quality inspection
• Learning Purified Feature Representations from Task-irrelevant Labels
• Direct Measure Matching for Crowd Counting
• Wasserstein Distance Regularized Sequence Representation for Text Matching in Asymmetrical Domains
• Multi-Purposing Domain Adaptation Discriminators for Pseudo Labeling Confidence
• Bi-Dimensional Feature Alignment for Cross-Domain Object Detection
• Detecting Bias in Transfer Learning Approaches for Text Classification
• Convergence of Gaussian-smoothed optimal transport distance with sub-gamma distributions and dependent samples
• Adaptive Wasserstein Hourglass for Weakly Supervised Hand Pose Estimation from Monocular RGB
• Unsupervised Domain Adaptation Network with Category-Centric Prototype Aligner for Biomedical Image Segmentation
• Sparsely-Labeled Source Assisted Domain Adaptation
• Information Theory-Guided Heuristic Progressive Multi-View Coding
• Zero-Round Active Learning
• KL Guided Domain Adaptation
• Joint Information Preservation for Heterogeneous Domain Adaptation
• Unsupervised Domain Adaptation for Cross-Subject Few-Shot Neurological Symptom Detection
• PyKale: Knowledge-Aware Machine Learning from Multiple Sources in Python
• Dual-Tuning: Joint Prototype Transfer and Structure Regularization for Compatible Feature Learning
• Exploring Dropout Discriminator for Domain Adaptation
• Teacher-Student Competition for Unsupervised Domain Adaptation
• Transferrable Contrastive Learning for Visual Domain Adaptation
• Towards All-around Knowledge Transferring: Learning From Task-irrelevant Labels
• Effective Unsupervised Domain Adaptation with Adversarially Trained Language Models
• The Role of Embedding Complexity in Domain-invariant Representations
• Learning on Random Balls is Sufficient for Estimating (Some) Graph Parameters
• Principled Representation Learning for Entity Alignment
• How does the Combined Risk Affect the Performance of Unsupervised Domain Adaptation Approaches?
• HRKD: Hierarchical Relational Knowledge Distillation for Cross-domain Language Model Compression
• Robust Ensembling Network for Unsupervised Domain Adaptation
• Open-Ended Visual Question Answering by Multi-Modal Domain Adaptation