Feature-Critic Networks for Heterogeneous Domain Generalization

Abstract

The well known domain shift issue causes model performance to degrade when deployed to a new target domain with different statistics to training. Domain adaptation techniques alleviate this, but need some instances from the target domain to drive adaptation. Domain generalisation is the recently topical problem of learning a model that generalises to unseen domains out of the box, and various approaches aim to train a domain-invariant feature extractor, typically by adding some manually designed losses. In this work, we propose a learning to learn approach, where the auxiliary loss that helps generalisation is itself learned. Beyond conventional domain generalisation, we consider a more challenging setting of heterogeneous domain generalisation, where the unseen domains do not share label space with the seen ones, and the goal is to train a feature representation that is useful off-the-shelf for novel data and novel categories. Experimental evaluation demonstrates that our method outperforms state-of-the-art solutions in both settings.

Presented at ICML 2019

References in corpus (9)

• Prototypical Networks for Few-shot Learning
• DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
• How transferable are features in deep neural networks?
• On the Convergence of Adam and Beyond
• Unsupervised Domain Adaptation with Residual Transfer Networks
• Meta-SGD: Learning to Learn Quickly for Few-Shot Learning
• Domain Separation Networks
• Domain Generalization via Invariant Feature Representation
• Learning to Learn: Meta-Critic Networks for Sample Efficient Learning

Cited by in corpus (30)

• Towards Out-Of-Distribution Generalization: A Survey
• Heterogeneous Domain Generalization via Domain Mixup
• Cross-Domain Few-Shot Classification via Learned Feature-Wise Transformation
• Discriminative Adversarial Domain Generalization with Meta-learning based Cross-domain Validation
• FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space
• Episodic Training for Domain Generalization
• Nonlinear Invariant Risk Minimization: A Causal Approach
• Learning to Generate Novel Domains for Domain Generalization
• FedH2L: Federated Learning with Model and Statistical Heterogeneity
• Online Meta-Critic Learning for Off-Policy Actor-Critic Methods
• How to Train Your MAML to Excel in Few-Shot Classification
• Revisiting Meta-Learning as Supervised Learning
• Frustratingly Easy Person Re-Identification: Generalizing Person Re-ID in Practice
• Semi-Supervised Domain Generalization with Stochastic StyleMatch
• Learning to Generalize Unseen Domains via Memory-based Multi-Source Meta-Learning for Person Re-Identification
• A Fourier-based Framework for Domain Generalization
• Learning to Learn with Variational Information Bottleneck for Domain Generalization
• Open Domain Generalization with Domain-Augmented Meta-Learning
• Cross-Domain Few-Shot Classification via Adversarial Task Augmentation
• Model-Based Domain Generalization
• A Channel Coding Benchmark for Meta-Learning
• Generalizable Model-agnostic Semantic Segmentation via Target-specific Normalization
• Provable Adaptation across Multiway Domains via Representation Learning
• An Optimization-Based Meta-Learning Model for MRI Reconstruction with Diverse Dataset
• Learning Invariant Representations across Domains and Tasks
• VideoDG: Generalizing Temporal Relations in Videos to Novel Domains
• Explainability-aided Domain Generalization for Image Classification
• Towards Recognizing New Semantic Concepts in New Visual Domains
• Domain Agnostic Learning for Unbiased Authentication
• Deep Domain Generalization with Feature-norm Network