試験
通常
試験時間: 00:00問題時間: 00:00
問題一覧
問題 39 / 65
Which data augmentation strategy is most effective when real data are extremely scarce for a rare disease?
①Perform a large number of only basic augmentations, such as simple noise addition and rotation.
②Perform only self-supervised learning on unlabeled data.
③Combine quality-controlled synthetic data with real data to augment it, and have the synthetic data validated by experts.
④Rely entirely on synthetic data and not use any real data.
⑤Only perform oversampling by duplicating existing rare disease cases.