Unsupervised Domain Adaptation through Self-Supervision

Abstract

This paper addresses unsupervised domain adaptation, the setting where labeled training data is available on a source domain, but the goal is to have good performance on a target domain with only unlabeled data. Like much of previous work, we seek to align the learned representations of the source and target domains while preserving discriminability. The way we accomplish alignment is by learning to perform auxiliary self-supervised task(s) on both domains simultaneously. Each self-supervised task brings the two domains closer together along the direction relevant to that task. Training this jointly with the main task classifier on the source domain is shown to successfully generalize to the unlabeled target domain. The presented objective is straightforward to implement and easy to optimize. We achieve state-of-the-art results on four out of seven standard benchmarks, and competitive results on segmentation adaptation. We also demonstrate that our method composes well with another popular pixel-level adaptation method.

References in corpus (7)

• Rethinking Atrous Convolution for Semantic Image Segmentation
• Understanding deep learning requires rethinking generalization
• CyCADA: Cycle-Consistent Adversarial Domain Adaptation
• Domain Separation Networks
• Unsupervised Cross-Domain Image Generation
• Domain Adaptation for Visual Applications: A Comprehensive Survey
• Boosting Supervision with Self-Supervision for Few-shot Learning

Cited by in corpus (27)

• Self-supervised Knowledge Distillation for Few-shot Learning
• Test-Time Adaptation to Distribution Shift by Confidence Maximization and Input Transformation
• Rethinking Distributional Matching Based Domain Adaptation
• Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation
• A Review of Single-Source Deep Unsupervised Visual Domain Adaptation
• Learning from Temporal Spatial Cubism for Cross-Dataset Skeleton-based Action Recognition
• Auxiliary Signal-Guided Knowledge Encoder-Decoder for Medical Report Generation
• Prototypical Cross-domain Self-supervised Learning for Few-shot Unsupervised Domain Adaptation
• Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization
• A Learnable Self-supervised Task for Unsupervised Domain Adaptation on Point Clouds
• Curriculum Manager for Source Selection in Multi-Source Domain Adaptation
• Graceful Degradation and Related Fields
• Cross-Domain Sentiment Classification with In-Domain Contrastive Learning
• Gradual Domain Adaptation in the Wild:When Intermediate Distributions are Absent
• Self-Supervised Euphemism Detection and Identification for Content Moderation
• Emotional Semantics-Preserved and Feature-Aligned CycleGAN for Visual Emotion Adaptation
• Curriculum CycleGAN for Textual Sentiment Domain Adaptation with Multiple Sources
• Unsupervised Multimodal Video-to-Video Translation via Self-Supervised Learning
• Multi-source Few-shot Domain Adaptation
• Multilevel Knowledge Transfer for Cross-Domain Object Detection
• Neural-Symbolic Solver for Math Word Problems with Auxiliary Tasks
• What Does Rotation Prediction Tell Us about Classifier Accuracy under Varying Testing Environments?
• An Efficient Method for the Classification of Croplands in Scarce-Label Regions
• RPCL: A Framework for Improving Cross-Domain Detection with Auxiliary Tasks
• Online Domain Adaptation for Continuous Cross-Subject Liver Viability Evaluation Based on Irregular Thermal Data
• Zero-Round Active Learning
• Test-Time Adaptation with Principal Component Analysis