%0 Conference Proceedings
%T Reducing Label Effort: Self-  Supervised Meets Active Learning
%A Javad Zolfaghari Bengar
%A Joost Van de Weijer
%A Bartlomiej Twardowski
%A Bogdan Raducanu
%B International Conference on Computer Vision Workshops
%D 2021
%F Javad Zolfaghari Bengar2021
%O LAMP;
%O exported from refbase (http://refbase.cvc.uab.es/show.php?record=3672), last updated on Mon, 15 Jul 2024 11:44:31 +0200
%X Active learning is a paradigm aimed at reducing the annotation effort by training the model on actively selected informative and/or representative samples. Another paradigm to reduce the annotation effort is self-training that learns from a large amount of unlabeled data in an unsupervised way and fine-tunes on few labeled samples. Recent developments in self-training have achieved very impressive results rivaling supervised learning on some datasets. The current work focuses on whether the two paradigms can benefit from each other. We studied object recognition datasets including CIFAR10, CIFAR100 and Tiny ImageNet with several labeling budgets for the evaluations. Our experiments reveal that self-training is remarkably more efficient than active learning at reducing the labeling effort, that for a low labeling budget, active learning offers no benefit to self-training, and finally that the combination of active learning and self-training is fruitful when the labeling budget is high. The performance gap between active learning trained either with self-training or from scratch diminishes as we approach to the point where almost half of the dataset is labeled.
%U https://ieeexplore.ieee.org/document/9607730
%U http://dx.doi.org/10.1109/ICCVW54120.2021.00188
%P 1631-1639