Learning to Generalize: Meta-Learning for Domain Generalization

Abstract

Domain shift refers to the well known problem that a model trained in one source domain performs poorly when applied to a target domain with different statistics. {Domain Generalization} (DG) techniques attempt to alleviate this issue by producing models which by design generalize well to novel testing domains. We propose a novel {meta-learning} method for domain generalization. Rather than designing a specific model that is robust to domain shift as in most previous DG work, we propose a model agnostic training procedure for DG. Our algorithm simulates train/test domain shift during training by synthesizing virtual testing domains within each mini-batch. The meta-optimization objective requires that steps to improve training domain performance should also improve testing domain performance. This meta-learning procedure trains models with good generalization ability to novel domains. We evaluate our method and achieve state of the art results on a recent cross-domain image classification benchmark, as well demonstrating its potential on two classic reinforcement learning tasks.

8 pages, 2 figures, under review of AAAI 2018

References in corpus (8)

• Frustratingly Easy Domain Adaptation
• Theoretical Models of Learning to Learn
• Matching Networks for One Shot Learning
• Learning multiple visual domains with residual adapters
• Universal representations:The missing link between faces, text, planktons, and cat breeds
• Learning to Learn: Meta-Critic Networks for Sample Efficient Learning
• Domain Generalization for Object Recognition with Multi-task Autoencoders
• Generalizing Skills with Semi-Supervised Reinforcement Learning

Cited by in corpus (21)

• One-Shot Imitation from Observing Humans via Domain-Adaptive Meta-Learning
• Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm
• DIVA: Domain Invariant Variational Autoencoders
• FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space
• Exploring Domain Shift in Extractive Text Summarization
• FSDR: Frequency Space Domain Randomization for Domain Generalization
• Domain Generalization Using a Mixture of Multiple Latent Domains
• Selecting Data Augmentation for Simulating Interventions
• Deep Stable Learning for Out-Of-Distribution Generalization
• Towards Shape Biased Unsupervised Representation Learning for Domain Generalization
• Joint Noise-Tolerant Learning and Meta Camera Shift Adaptation for Unsupervised Person Re-Identification
• POEM: Polarization of Embeddings for Domain-Invariant Representations
• Detecting, Localising and Classifying Polyps from Colonoscopy Videos using Deep Learning
• Meta Batch-Instance Normalization for Generalizable Person Re-Identification
• Domain2Vec: Deep Domain Generalization
• Robustness through Data Augmentation Loss Consistency
• Learning Image Labels On-the-fly for Training Robust Classification Models
• Coverage-Guaranteed Prediction Sets for Out-of-Distribution Data
• Contrastive ACE: Domain Generalization Through Alignment of Causal Mechanisms
• Meta Gradient Adversarial Attack
• Neuron Coverage-Guided Domain Generalization