Cross-Domain Few-Shot Classification via Adversarial Task Augmentation

2021/04/29 by Haoqing Wang, Wang, Haoqing, Zhihong Deng +2 · 9 citations
Computer Science · Mathematics · Medicine · #Adversarial system #Artificial intelligence #COVID-19 diagnosis using AI #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Domain (mathematical analysis) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Generalization #Inductive bias #Machine learning #Mathematics #Meta learning (computer science) #Multi-task learning #Robustness (evolution) #Task (project management) #cs.CV

paper · pdf · doi:10.48550/arxiv.2104.14385

published in arXiv (Cornell University) (Cornell University) · Accepted by IJCAI-21 (the 30th International Joint Conference on Artificial Intelligence) Main Track

openalex publication_date 2021/04/29 · arxiv created 2021/05/02 · arxiv updated 2021/05/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/06

Abstract

Few-shot classification aims to recognize unseen classes with few labeled samples from each class. Many meta-learning models for few-shot classification elaborately design various task-shared inductive bias (meta-knowledge) to solve such tasks, and achieve impressive performance. However, when there exists the domain shift between the training tasks and the test tasks, the obtained inductive bias fails to generalize across domains, which degrades the performance of the meta-learning models. In this work, we aim to improve the robustness of the inductive bias through task augmentation. Concretely, we consider the worst-case problem around the source task distribution, and propose the adversarial task augmentation method which can generate the inductive bias-adaptive 'challenging' tasks. Our method can be used as a simple plug-and-play module for various meta-learning models, and improve their cross-domain generalization capability. We conduct extensive experiments under the cross-domain setting, using nine few-shot classification datasets: mini-ImageNet, CUB, Cars, Places, Plantae, CropDiseases, EuroSAT, ISIC and ChestX. Experimental results show that our method can effectively improve the few-shot classification performance of the meta-learning models under domain shift, and outperforms the existing works. Our code is available at https://github.com/Haoqing-Wang/CDFSL-ATA.

Citations

• Human-level concept learning through probabilistic program induction
• Learning to Propagate Labels: Transductive Propagation Network for Few-shot Learning
• Feature Transformation Ensemble Model with Batch Spectral Regularization for Cross-Domain Few-Shot Classification

Cited by

• Free Lunch to Meet the Gap: Intermediate Domain Reconstruction for Cross-Domain Few-Shot Learning
• FreqGRL: Suppressing Low-Frequency Bias and Mining High-Frequency Knowledge for Cross-Domain Few-Shot Learning
• Source-Free Cross-Domain Continual Learning
• Cross-Domain Few-Shot Learning with Coalescent Projections and Latent Space Reservation
• Learning to Adapt Frozen CLIP for Few-Shot Test-Time Domain Adaptation
• Random Registers for Cross-Domain Few-Shot Learning
• Revisiting Continuity of Image Tokens for Cross-domain Few-shot Learning
• Enhance Then Search: An Augmentation-Search Strategy with Foundation Models for Cross-Domain Few-Shot Object Detection
• A Cross-Domain Few-Shot Learning Method Based on Domain Knowledge Mapping

Related

• Adversarial Feature Augmentation for Cross-domain Few-shot Classification
• Domain-Agnostic Few-Shot Classification by Learning Disparate Modulators
• Cross-domain Few-shot Learning with Task-specific Adapters
• Combining Domain-Specific Meta-Learners in the Parameter Space for Cross-Domain Few-Shot Classification
• Task-Level Contrastiveness for Cross-Domain Few-Shot Learning