Outperforms state-of-the-art algorithms in deep learning tasks

Newswise — Deep learning based semi-supervised learning algorithms have shown promising results in recent years. Their Meta-Semi approach outperforms state-of-the-art semi-supervised learning algorithms. “Most deep learning based semi-supervised learning algorithms introduce multiple tunable hyper-parameters, making them less practical in real semi-supervised learning scenarios where the labeled data is scarce for extensive hyper-parameter search,” said Huang. “Empirically, Meta-Semi outperforms state-of-the-art semi-supervised learning algorithms significantly on the challenging semi-supervised CIFAR-100 and STL-10 tasks, and achieves competitive performance on CIFAR-10 and SVHN,” said Huang. Compared to existing deep semi-supervised learning algorithms, Meta-Semi requires much less effort for tuning hyper-parameters, but achieves state-of-the-art performance on the four competitive datasets.