<ul><li>Proposed evidential deep active learning approach for semi-supervised classification (EDALSSC) focuses on uncertainty estimation of prediction results during the learning process.</li><li>EDALSSC builds a framework to quantify uncertainty estimation of labeled and unlabeled data simultaneously using evidential deep learning.</li><li>The uncertainty estimation of labeled data involves evidential deep learning, while that of unlabeled data is modeled by combining ignorance and conflict information of evidence.</li><li>Experimental results show that EDALSSC outperforms existing semi-supervised and supervised active learning approaches on image classification datasets.</li></ul>

Evidential Deep Active Learning for Semi-Supervised Classification

Discover more