Papers with self-training algorithms
Neural Self-Training through Spaced Repetition (N19-1)
Copied to clipboard
| Challenge: | Existing methods for self-training rely on predetermined policies to sample unlabeled data. |
| Approach: | They propose a semi-supervised learning approach that uses spaced repetition to dynamically sample informative and diverse unlabeled instances with respect to individual learner and instance characteristics. |
| Outcome: | The proposed model outperforms existing semi-supervised learning approaches on publicly-available datasets. |