Unsupervised Domain Adaptation via Calibrating Uncertainties

Abstract

Unsupervised domain adaptation (UDA) aims at inferring class labels for unlabeled target domain given a related labeled source dataset. Intuitively, a model trained on source domain normally produces higher uncertainties for unseen data. In this work, we build on this assumption and propose to adapt from source to target domain via calibrating their predictive uncertainties. The uncertainty is quantified as the Renyi entropy, from which we propose a general Renyi entropy regularization (RER) framework. We further employ variational Bayes learning for reliable uncertainty estimation. In addition, calibrating the sample variance of network parameters serves as a plug-in regularizer for training. We discuss the theoretical properties of the proposed method and demonstrate its effectiveness on three domain-adaptation tasks.

arXiv comment4 pages

Cited by in corpus (10)

• Addressing Failure Prediction by Learning Model Confidence
• Uncertainty-Aware Consistency Regularization for Cross-Domain Semantic Segmentation
• Confidence Regularized Self-Training
• Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation
• Conservative Wasserstein Training for Pose Estimation
• Synergizing between Self-Training and Adversarial Learning for Domain Adaptive Object Detection
• Uncertainty-Aware Model Adaptation for Unsupervised Cross-Domain Object Detection
• Energy-constrained Self-training for Unsupervised Domain Adaptation
• Generative Self-training for Cross-domain Unsupervised Tagged-to-Cine MRI Synthesis
• Certainty Volume Prediction for Unsupervised Domain Adaptation