ANEDL: Adaptive Negative Evidential Deep Learning for Open-Set Semi-supervised Learning
Yang Yu, Danruo Deng, Furui Liu, Qi Dou, Yueming Jin, Guangyong Chen, Pheng Ann Heng
Abstract
Semi-supervised learning (SSL) methods assume that labeled data, unlabeled data and test data are from the same distribution. Open-set semi-supervised learning (Open-set SSL) con- siders a more practical scenario, where unlabeled data and test data contain new categories (outliers) not observed in labeled data (inliers). Most previous works focused on out- lier detection via binary classifiers, which suffer from insufficient scalability and inability to distinguish different types of uncertainty. In this paper, we propose a novel framework, Adaptive Negative Evidential Deep Learning (ANEDL) to tackle these limitations. Concretely, we first introduce evidential deep learning (EDL) as an outlier detector to quantify different types of uncertainty, and design different uncertainty metrics for self-training and inference. Furthermore, we propose a novel adaptive negative optimization strategy, making EDL more tailored to the unlabeled dataset containing both inliers and outliers. As demonstrated empirically, our proposed method outperforms existing state-of-the-art methods across four datasets.
BibTeX
@article{Yu_Deng_Liu_Dou_Jin_Chen_Heng_2024, title={ANEDL: Adaptive Negative Evidential Deep Learning for Open-Set Semi-supervised Learning}, volume={38}, url={https://ojs.aaai.org/index.php/AAAI/article/view/29597}, DOI={10.1609/aaai.v38i15.29597}, abstractNote={Semi-supervised learning (SSL) methods assume that labeled
data, unlabeled data and test data are from the same distribution. Open-set semi-supervised learning (Open-set SSL) con-
siders a more practical scenario, where unlabeled data and
test data contain new categories (outliers) not observed in
labeled data (inliers). Most previous works focused on out-
lier detection via binary classifiers, which suffer from insufficient scalability and inability to distinguish different types
of uncertainty. In this paper, we propose a novel framework,
Adaptive Negative Evidential Deep Learning (ANEDL) to
tackle these limitations. Concretely, we first introduce evidential deep learning (EDL) as an outlier detector to quantify
different types of uncertainty, and design different uncertainty
metrics for self-training and inference. Furthermore, we propose a novel adaptive negative optimization strategy, making
EDL more tailored to the unlabeled dataset containing both
inliers and outliers. As demonstrated empirically, our proposed method outperforms existing state-of-the-art methods
across four datasets.}, number={15}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Yu, Yang and Deng, Danruo and Liu, Furui and Dou, Qi and Jin, Yueming and Chen, Guangyong and Heng, Pheng Ann}, year={2024}, month={Mar.}, pages={16587-16595} }