Domain Generalization via Invariant Feature Representation

Abstract

This paper investigates domain generalization: How to take knowledge acquired from an arbitrary number of related domains and apply it to previously unseen domains? We propose Domain-Invariant Component Analysis (DICA), a kernel-based optimization algorithm that learns an invariant transformation by minimizing the dissimilarity across domains, whilst preserving the functional relationship between input and output variables. A learning-theoretic analysis shows that reducing dissimilarity improves the expected generalization ability of classifiers on new domains, motivating the proposed algorithm. Experimental results on synthetic and real-world datasets demonstrate that DICA successfully learns invariant features and improves classifier performance in practice.

The 30th International Conference on Machine Learning (ICML 2013)

References in corpus (2)

• Learning from Distributions via Support Measure Machines
• Flexible Modeling of Latent Task Structures in Multitask Learning

Cited by in corpus (183)

• A review of domain adaptation without target labels
• Beyond Sharing Weights for Deep Domain Adaptation
• Kernel Mean Embedding of Distributions: A Review and Beyond
• Learning Decentralized Controllers for Robot Swarms with Graph Neural Networks
• Generalizing to Unseen Domains via Adversarial Data Augmentation
• Generalizing Across Domains via Cross-Gradient Training
• Patch-based Output Space Adversarial Learning for Joint Optic Disc and Cup Segmentation
• WILDS: A Benchmark of in-the-Wild Distribution Shifts
• Learning Robust Global Representations by Penalizing Local Predictive Power
• An introduction to domain adaptation and transfer learning
• Towards Out-Of-Distribution Generalization: A Survey
• Measuring Robustness to Natural Distribution Shifts in Image Classification
• Invariant Models for Causal Transfer Learning
• Domain Generalization by Marginal Transfer Learning
• Heterogeneous Domain Generalization via Domain Mixup
• Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation
• Causally Regularized Learning with Agnostic Data Selection Bias
• The Risks of Invariant Risk Minimization
• DIVA: Domain Invariant Variational Autoencoders
• Generalizing to unseen domains via distribution matching
• MEMO: Test Time Robustness via Adaptation and Augmentation
• In Search of Lost Domain Generalization
• Domain Generalization for Medical Imaging Classification with Linear-Dependency Regularization
• Domain Generalization using Causal Matching
• Feature-Critic Networks for Heterogeneous Domain Generalization
• Test-Time Training with Self-Supervision for Generalization under Distribution Shifts
• Learning explanations that are hard to vary
• Domain Generalization via Optimal Transport with Metric Similarity Learning
• Efficient Domain Generalization via Common-Specific Low-Rank Decomposition
• Frustratingly Simple Domain Generalization via Image Stylization
• Understanding the Failure Modes of Out-of-Distribution Generalization
• Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
• Domain Generalization via Multidomain Discriminant Analysis
• Ensemble of Averages: Improving Model Selection and Boosting Performance in Domain Generalization
• Learning Causal Semantic Representation for Out-of-Distribution Prediction
• Patchwork Learning: A Paradigm Towards Integrative Analysis across Diverse Biomedical Data Sources
• Invariance Principle Meets Information Bottleneck for Out-of-Distribution Generalization
• Self-Challenging Improves Cross-Domain Generalization
• Support and Invertibility in Domain-Invariant Representations
• Towards a Theoretical Framework of Out-of-Distribution Generalization
• AdaRNN: Adaptive Learning and Forecasting of Time Series
• Inferring Latent Domains for Unsupervised Deep Domain Adaptation
• Feature Alignment and Restoration for Domain Generalization and Adaptation
• Learning Attributes Equals Multi-Source Domain Generalization
• Learning to Optimize Domain Specific Normalization for Domain Generalization
• FSDR: Frequency Space Domain Randomization for Domain Generalization
• Federated Learning with Domain Generalization
• Unseen Target Stance Detection with Adversarial Domain Generalization
• Episodic Training for Domain Generalization
• Domain Adaptation as a Problem of Inference on Graphical Models
• Nonlinear Invariant Risk Minimization: A Causal Approach
• Learning to Generate Novel Domains for Domain Generalization
• A Review of Single-Source Deep Unsupervised Visual Domain Adaptation
• Towards Non-I.I.D. Image Classification: A Dataset and Baselines
• Adversarial Invariant Feature Learning with Accuracy Constraint for Domain Generalization
• Style Normalization and Restitution for Generalizable Person Re-identification
• Deep Stable Learning for Out-Of-Distribution Generalization
• Selecting Data Augmentation for Simulating Interventions
• Domain Randomization and Pyramid Consistency: Simulation-to-Real Generalization without Accessing Target Domain Data
• BREEDS: Benchmarks for Subpopulation Shift
• Transfer Metric Learning: Algorithms, Applications and Outlooks
• Contrastive Syn-to-Real Generalization
• Can Subnetwork Structure be the Key to Out-of-Distribution Generalization?
• Zero-shot Domain Adaptation without Domain Semantic Descriptors
• Empirical or Invariant Risk Minimization? A Sample Complexity Perspective
• GraphGLOW: Universal and Generalizable Structure Learning for Graph Neural Networks
• Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond
• MixStyle Neural Networks for Domain Generalization and Adaptation
• Multi-Scale and Multi-Layer Contrastive Learning for Domain Generalization
• Learning to Generalize Unseen Domains via Memory-based Multi-Source Meta-Learning for Person Re-Identification
• Towards Universal Representation Learning for Deep Face Recognition
• Domain adaptation under structural causal models
• Multiple Domain Experts Collaborative Learning: Multi-Source Domain Generalization For Person Re-Identification
• Progressive Domain Expansion Network for Single Domain Generalization
• EllSeg-Gen, towards Domain Generalization for head-mounted eyetracking
• Variational Disentanglement for Domain Generalization
• Generalizable Person Re-identification with Relevance-aware Mixture of Experts
• Representation via Representations: Domain Generalization via Adversarially Learned Invariant Representations
• A Fourier-based Framework for Domain Generalization
• Learning to Learn with Variational Information Bottleneck for Domain Generalization
• Domain Invariant Representation Learning with Domain Density Transformations
• Fundamental Limits and Tradeoffs in Invariant Representation Learning
• On the Transfer of Disentangled Representations in Realistic Settings
• Towards Open-World Feature Extrapolation: An Inductive Graph Learning Approach
• Few-Shot Learning as Domain Adaptation: Algorithm and Analysis
• Anomalous Example Detection in Deep Learning: A Survey
• Kernel Mean Estimation via Spectral Filtering
• Open Domain Generalization with Domain-Augmented Meta-Learning
• Learning Domain Invariant Representations by Joint Wasserstein Distance Minimization
• Style Normalization and Restitution for Domain Generalization and Adaptation
• Respecting Domain Relations: Hypothesis Invariance for Domain Generalization
• Enforcing Predictive Invariance across Structured Biomedical Domains
• Learning to Learn Single Domain Generalization
• I-SPEC: An End-to-End Framework for Learning Transportable, Shift-Stable Models
• Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Tests
• Deep Domain-Adversarial Image Generation for Domain Generalisation
• Improve Model Generalization and Robustness to Dataset Bias with Bias-regularized Learning and Domain-guided Augmentation
• Few-Shot Domain Expansion for Face Anti-Spoofing
• Coping with Label Shift via Distributionally Robust Optimisation
• A Theory of Multiple-Source Adaptation with Limited Target Labeled Data
• MetaAlign: Coordinating Domain Alignment and Classification for Unsupervised Domain Adaptation
• Model-Based Domain Generalization
• An Online Learning Approach to Interpolation and Extrapolation in Domain Generalization
• Quantifying and Improving Transferability in Domain Generalization
• Learning Meta Face Recognition in Unseen Domains
• Cross-Domain First Person Audio-Visual Action Recognition through Relative Norm Alignment
• Learn to Expect the Unexpected: Probably Approximately Correct Domain Generalization
• A Discriminative Technique for Multiple-Source Adaptation
• Stable Prediction with Model Misspecification and Agnostic Distribution Shift
• A Bit More Bayesian: Domain-Invariant Learning with Uncertainty
• Environment Invariant Linear Least Squares
• Affective Image Content Analysis: Two Decades Review and New Perspectives
• Novel Human-Object Interaction Detection via Adversarial Domain Generalization
• Domain Generalization via Inference-time Label-Preserving Target Projections
• Learning to Diversify for Single Domain Generalization
• Towards Principled Disentanglement for Domain Generalization
• Improving Confidence Estimates for Unfamiliar Examples
• Iterative Feature Matching: Toward Provable Domain Generalization with Logarithmic Environments
• Domain2Vec: Deep Domain Generalization
• Moment-Based Domain Adaptation: Learning Bounds and Algorithms
• WEDGE: Web-Image Assisted Domain Generalization for Semantic Segmentation
• Bilevel Online Adaptation for Out-of-Domain Human Mesh Reconstruction
• Class-conditioned Domain Generalization via Wasserstein Distributional Robust Optimization
• Domain Generalization by Functional Regression
• SWAD: Domain Generalization by Seeking Flat Minima
• Transfer String Kernel for Cross-Context DNA-Protein Binding Prediction
• Model-Based Robust Deep Learning: Generalizing to Natural, Out-of-Distribution Data
• Causality-inspired Single-source Domain Generalization for Medical Image Segmentation
• S2R-DepthNet: Learning a Generalizable Depth-specific Structural Representation
• An Information-theoretic Approach to Distribution Shifts
• The Connection between Out-of-Distribution Generalization and Privacy of ML Models
• Global Convergence and Generalization Bound of Gradient-Based Meta-Learning with Deep Neural Nets
• Generalizable Model-agnostic Semantic Segmentation via Target-specific Normalization
• Uncertainty-guided Model Generalization to Unseen Domains
• Diagnosing the Environment Bias in Vision-and-Language Navigation
• Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net
• Image Alignment in Unseen Domains via Domain Deep Generalization
• Hierarchical Variational Auto-Encoding for Unsupervised Domain Generalization
• Sequential Learning for Domain Generalization
• Deep Unified Representation for Heterogeneous Recommendation
• Conditional Coupled Generative Adversarial Networks for Zero-Shot Domain Adaptation
• DLOW: Domain Flow for Adaptation and Generalization
• Embracing the Dark Knowledge: Domain Generalization Using Regularized Knowledge Distillation
• Contrastive ACE: Domain Generalization Through Alignment of Causal Mechanisms
• A Riemannian gossip approach to subspace learning on Grassmann manifold
• Towards Recognizing New Semantic Concepts in New Visual Domains
• Two Sides of Meta-Learning Evaluation: In vs. Out of Distribution
• Weighed Domain-Invariant Representation Learning for Cross-domain Sentiment Analysis
• Explainability-aided Domain Generalization for Image Classification
• Neural Networks for Learning Counterfactual G-Invariances from Single Environments
• Unsupervised Domain Expansion from Multiple Sources
• Discrepancy Minimization in Domain Generalization with Generative Nearest Neighbors
• Adversarial Learning for Zero-shot Domain Adaptation
• Contrastive Language Adaptation for Cross-Lingual Stance Detection
• Deep Transfer Learning for Thermal Dynamics Modeling in Smart Buildings
• Privacy-Preserving Constrained Domain Generalization via Gradient Alignment
• Kernel Embedding Approaches to Orbit Determination of Spacecraft Clusters
• Confounder Identification-free Causal Visual Feature Learning
• Deep Shape Matching
• Learning Generalized Spatial-Temporal Deep Feature Representation for No-Reference Video Quality Assessment
• KL Guided Domain Adaptation
• Balance-Subsampled Stable Prediction
• Domain Generalization via Domain-based Covariance Minimization
• Discriminative Domain-Invariant Adversarial Network for Deep Domain Generalization
• Distributionally Robust Learning with Stable Adversarial Training
• Discovering Spatial Relationships by Transformers for Domain Generalization
• Learning to see across Domains and Modalities
• Effect-Invariant Mechanisms for Policy Generalization
• Stable Prediction across Unknown Environments
• Boosting Unconstrained Face Recognition with Auxiliary Unlabeled Data
• Domain Agnostic Learning for Unbiased Authentication
• Learning to Generalize One Sample at a Time with Self-Supervision
• Reappraising Domain Generalization in Neural Networks
• Collaborating Foundation Models for Domain Generalized Semantic Segmentation
• Neuron Coverage-Guided Domain Generalization
• Causality and Generalizability: Identifiability and Learning Methods
• Towards Data-Free Domain Generalization
• Achieving Domain Robustness in Stereo Matching Networks by Removing Shortcut Learning
• Global and Local Texture Randomization for Synthetic-to-Real Semantic Segmentation
• Deep Domain Generalization with Feature-norm Network
• Counterfactual Supervision-based Information Bottleneck for Out-of-Distribution Generalization
• Domain Generalization via Universal Non-volume Preserving Models
• Exploring Data Aggregation and Transformations to Generalize across Visual Domains