学习算法的成功取决于其在拟合数据和避免过拟合之间取得平衡的能力。
The success of a learning algorithm depends on its ability to balance between fitting the data and avoiding overfitting.
The only way to achieve the impossible is to believe it is possible.