Learning to Generate Novel Domains for Domain Generalization

2020/07/07 by Kaiyang Zhou, Yongxin Yang, Zhou, Kaiyang +6 · 31 citations
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Multimodal Machine Learning Applications #Topic Modeling #cs.CV

paper · pdf · doi:10.48550/arxiv.2007.03304

ECCV'20

openalex publication_date 2020/07/07 · arxiv created 2021/03/09 · arxiv updated 2021/03/10 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper focuses on domain generalization (DG), the task of learning from multiple source domains a model that generalizes well to unseen domains. A main challenge for DG is that the available source domains often exhibit limited diversity, hampering the model's ability to learn to generalize. We therefore employ a data generator to synthesize data from pseudo-novel domains to augment the source domains. This explicitly increases the diversity of available training domains and leads to a more generalizable model. To train the generator, we model the distribution divergence between source and synthesized pseudo-novel domains using optimal transport, and maximize the divergence. To ensure that semantics are preserved in the synthesized data, we further impose cycle-consistency and classification losses on the generator. Our method, L2A-OT (Learning to Augment by Optimal Transport) outperforms current state-of-the-art DG methods on four benchmark datasets.

Citations

• A unifying view on dataset shift in classification
• Gradient-based learning applied to document recognition
• Domain Generalization via Invariant Feature Representation
• Learning Transferable Features with Deep Adaptation Networks
• The Cramer Distance as a Solution to Biased Wasserstein Gradients
• Domain Generalization via Model-Agnostic Learning of Semantic Features

Cited by

• RobustSurg-Tackling domain generalisation for out-of-distribution surgical scene segmentation
• The Finer the Better: Towards Granular-aware Open-set Domain Generalization
• D-GAP: Improving Out-of-Domain Robustness via Dataset-Agnostic and Gradient-Guided Augmentation in Amplitude and Pixel Spaces
• A Dual-stage Prompt-driven Privacy-preserving Paradigm for Person Re-Identification
• Humanoid-inspired Causal Representation Learning for Domain Generalization
• Connecting Domains and Contrasting Samples: A Ladder for Domain Generalization
• Domain Generalizable Continual Learning
• Measure-Theoretic Anti-Causal Representation Learning
• EReLiFM: Evidential Reliability-Aware Residual Flow Meta-Learning for Open-Set Domain Generalization under Noisy Labels
• CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization
• Single Domain Generalization in Diabetic Retinopathy: A Neuro-Symbolic Learning Approach
• Feed Two Birds with One Scone: Exploiting Function-Space Regularization for Both OOD Robustness and ID Fine-Tuning Performance
• GNN-based Unified Deep Learning
• Simulate, Refocus and Ensemble: An Attention-Refocusing Scheme for Domain Generalization
• FEDTAIL: Federated Long-Tailed Domain Generalization with Sharpness-Guided Gradient Matching
• FedPall: Prototype-based Adversarial and Collaborative Learning for Federated Learning with Feature Drift
• Generalizing vision-language models to novel domains: A comprehensive survey
• Learning Time-Aware Causal Representation for Model Generalization in Evolving Domains
• Learning to Adapt Frozen CLIP for Few-Shot Test-Time Domain Adaptation
• Optimal Transport-based Domain Alignment as a Preprocessing Step for Federated Learning
• RCCDA: Adaptive Model Updates in the Presence of Concept Drift under a Constrained Resource Budget
• Pseudo Multi-Source Domain Generalization: Bridging the Gap Between Single and Multi-Source Domain Generalization
• LangDAug: Langevin Data Augmentation for Multi-Source Domain Generalization in Medical Image Segmentation
• Multi-Source Collaborative Style Augmentation and Domain-Invariant Learning for Federated Domain Generalization
• A Bit More Bayesian: Domain-Invariant Learning with Uncertainty
• Neuron Coverage-Guided Domain Generalization
• TestDG: Test-time Domain Generalization for Continual Test-time Adaptation
• CLIP-Powered Domain Generalization and Domain Adaptation: A Comprehensive Survey
• Stratify: Rethinking Federated Learning for Non-IID Data through Balanced Sampling
• Vision and Language Integration for Domain Generalization
• DGFamba: Learning Flow Factorized State Space for Visual Domain Generalization

Related

• In Search of Lost Domain Generalization
• Learning to Learn Single Domain Generalization
• Learning to Diversify for Single Domain Generalization
• Generalizing to Unseen Domains: A Survey on Domain Generalization
• Towards Unsupervised Domain Generalization