写博客初衷
我在看机器学习算法过程中一个发现就是：往往弄清一个算法是什么（或者再加一个干什么的？）并不复杂，算法本身的过程并不复杂，甚至可以说非常简单；但如果你要深究算法每一步是为什么，或者说寻求一个（或数个，神奇的是常常数个不同的解释最后殊途同归导向同一个算法过程）合理的解释的过程就比较曲折了。
I found a phenomena when I was learning how machine is learning: it is usually not too difficult to figure out what a algorithm is and do, sometimes it is even fairly easy. But if you want to know why and how it do this, that is you want find a interpretation about why and how it works, things can be very complex. Machine learning algorithms are hard to understand in the sense that it's hard to understand the interpretation that justifies the learning procedure. Human are curious animals, we are not satisfied with knowing a series magic procedure without knowing the source of magic. It certainly should not be blamed for our curiosities. But it's necessary to appreciate the magic before understanding it. This may be a natural and sensible learning order, don't try to understand why try to figure out what happens and get some intuitive impressions at the first step of learning a algorithm.