深度学习模型经常受限于小数据集，我们可以使用数据增强 (Data Augmentation) 技术来生成更大规模更具多样性的训练集，从而提升模型的鲁棒性和性能。
Deep Learning models are data-greedy and often limited by small datasets. Data Augmentation techniques can be applied to boost training datasets, so that the performance and robustness of models may be improved.
