Domain Generalization via Model-Agnostic Learning of Semantic Features

Abstract

Generalization capability to unseen domains is crucial for machine learning models when deploying to real-world conditions. We investigate the challenging problem of domain generalization, i.e., training a model on multi-domain source data such that it can directly generalize to target domains with unknown statistics. We adopt a model-agnostic learning paradigm with gradient-based meta-train and meta-test procedures to expose the optimization to domain shift. Further, we introduce two complementary losses which explicitly regularize the semantic structure of the feature space. Globally, we align a derived soft confusion matrix to preserve general knowledge about inter-class relationships. Locally, we promote domain-independent class-specific cohesion and separation of sample features with a metric-learning component. The effectiveness of our method is demonstrated with new state-of-the-art results on two common object recognition benchmarks. Our method also shows consistent improvement on a medical image segmentation task.

arXiv commentNeurIPS 2019

Cited by in corpus (92)

• MedViT: A Robust Vision Transformer for Generalized Medical Image Classification
• WILDS: A Benchmark of in-the-Wild Distribution Shifts
• Towards Data-Efficient Learning: A Benchmark for COVID-19 CT Lung and Infection Segmentation
• Towards Out-Of-Distribution Generalization: A Survey
• Test-Time Adaptable Neural Networks for Robust Medical Image Segmentation
• AADG: Automatic Augmentation for Domain Generalization on Retinal Image Segmentation
• Generalizing to unseen domains via distribution matching
• In Search of Lost Domain Generalization
• Domain Generalization using Causal Matching
• Domain Generalization via Optimal Transport with Metric Similarity Learning
• Efficient Domain Generalization via Common-Specific Low-Rank Decomposition
• SAN-Net: Learning Generalization to Unseen Sites for Stroke Lesion Segmentation with Self-Adaptive Normalization
• Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety
• FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space
• Domain Generalization with MixStyle
• Improving Generalizability of Graph Anomaly Detection Models via Data Augmentation
• Auto-FedAvg: Learnable Federated Averaging for Multi-Institutional Medical Image Segmentation
• Learning to Optimize Domain Specific Normalization for Domain Generalization
• FSDR: Frequency Space Domain Randomization for Domain Generalization
• Automated Claustrum Segmentation in Human Brain MRI Using Deep Learning
• Improving out-of-distribution generalization via multi-task self-supervised pretraining
• Nonlinear Invariant Risk Minimization: A Causal Approach
• Learning to Generate Novel Domains for Domain Generalization
• Deep Stable Learning for Out-Of-Distribution Generalization
• MLN-net: A multi-source medical image segmentation method for clustered microcalcifications using multiple layer normalization
• Self-Supervision with Superpixels: Training Few-shot Medical Image Segmentation without Annotation
• Domain Adaptive Relational Reasoning for 3D Multi-Organ Segmentation
• Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization
• Gradient Matching for Domain Generalization
• Progressive Domain Expansion Network for Single Domain Generalization
• POEM: Polarization of Embeddings for Domain-Invariant Representations
• Multiple Domain Experts Collaborative Learning: Multi-Source Domain Generalization For Person Re-Identification
• Shape-aware Meta-learning for Generalizing Prostate MRI Segmentation to Unseen Domains
• Variational Disentanglement for Domain Generalization
• Learning to Learn with Variational Information Bottleneck for Domain Generalization
• A Fourier-based Framework for Domain Generalization
• Generalizable Person Re-identification with Relevance-aware Mixture of Experts
• Unsupervised Bidirectional Cross-Modality Adaptation via Deeply Synergistic Image and Feature Alignment for Medical Image Segmentation
• Delving Deep into the Generalization of Vision Transformers under Distribution Shifts
• Magnification Generalization for Histopathology Image Embedding
• Enforcing Predictive Invariance across Structured Biomedical Domains
• Learning Domain Invariant Representations by Joint Wasserstein Distance Minimization
• Adaptive Domain Generalization via Online Disagreement Minimization
• Open Domain Generalization with Domain-Augmented Meta-Learning
• Learning Cross-domain Generalizable Features by Representation Disentanglement
• Parameter Exchange for Robust Dynamic Domain Generalization
• Online Meta-Learning for Multi-Source and Semi-Supervised Domain Adaptation
• Domain Generalization via Semi-supervised Meta Learning
• Cross-Domain First Person Audio-Visual Action Recognition through Relative Norm Alignment
• Domain Generalization for Session-Independent Brain-Computer Interface
• Towards Domain Invariant Single Image Dehazing
• Model-Based Domain Generalization
• MetaAlign: Coordinating Domain Alignment and Classification for Unsupervised Domain Adaptation
• Learn to Expect the Unexpected: Probably Approximately Correct Domain Generalization
• Re-using Adversarial Mask Discriminators for Test-time Training under Distribution Shifts
• A Bit More Bayesian: Domain-Invariant Learning with Uncertainty
• AtrialGeneral: Domain Generalization for Left Atrial Segmentation of Multi-Center LGE MRIs
• Unsupervised Cross-domain Image Classification by Distance Metric Guided Feature Alignment
• Learning to Diversify for Single Domain Generalization
• Feature Stylization and Domain-aware Contrastive Learning for Domain Generalization
• SWAD: Domain Generalization by Seeking Flat Minima
• Domain Generalization via Inference-time Label-Preserving Target Projections
• Self-balanced Learning For Domain Generalization
• Causality-inspired Single-source Domain Generalization for Medical Image Segmentation
• Barycentric-alignment and reconstruction loss minimization for domain generalization
• Uncertainty-guided Model Generalization to Unseen Domains
• Domain-Invariant Speaker Vector Projection by Model-Agnostic Meta-Learning
• Learning Image Labels On-the-fly for Training Robust Classification Models
• Random Style Transfer based Domain Generalization Networks Integrating Shape and Spatial Information
• Cooperative Training and Latent Space Data Augmentation for Robust Medical Image Segmentation
• Style-invariant Cardiac Image Segmentation with Test-time Augmentation
• Embracing the Dark Knowledge: Domain Generalization Using Regularized Knowledge Distillation
• Text is Text, No Matter What: Unifying Text Recognition using Knowledge Distillation
• Multi-task Learning by Leveraging the Semantic Information
• VideoDG: Generalizing Temporal Relations in Videos to Novel Domains
• Privacy-Preserving Constrained Domain Generalization via Gradient Alignment
• Pareto Navigation Gradient Descent: a First-Order Algorithm for Optimization in Pareto Set
• Confounder Identification-free Causal Visual Feature Learning
• Collaborative Semantic Aggregation and Calibration for Federated Domain Generalization
• Histogram Matching Augmentation for Domain Adaptation with Application to Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Image Segmentation
• Explainability-aided Domain Generalization for Image Classification
• Discrepancy Minimization in Domain Generalization with Generative Nearest Neighbors
• Contrastive ACE: Domain Generalization Through Alignment of Causal Mechanisms
• Mutual Information-based Disentangled Neural Networks for Classifying Unseen Categories in Different Domains: Application to Fetal Ultrasound Imaging
• Deep Domain Generalization with Feature-norm Network
• Neuron Coverage-Guided Domain Generalization
• An Empirical Framework for Domain Generalization in Clinical Settings
• Domain Generalization for Mammography Detection via Multi-style and Multi-view Contrastive Learning
• Robust Domain-Free Domain Generalization with Class-aware Alignment
• Discovering Spatial Relationships by Transformers for Domain Generalization
• Towards Data-Free Domain Generalization
• Domain Composition and Attention for Unseen-Domain Generalizable Medical Image Segmentation