In machine learning, the trade-off between bias and variance is crucial for achieving good generalization.
通过经验获得的东西比通过推理获得的东西更真实、更坚实。
提高的唯一方法就是与最强者对抗。