Machine learning---"No free Lunch" (no lunch) theorem easy to understand explanation _ depth learning/machine learning

Students in the field of machine learning know that there is a universal theorem in machine learning: There is no free lunch (no lunch).

The simple and understandable explanation for it is this:

1, an algorithm (algorithm a) on a specific data set than the performance of another algorithm (algorithm B) at the same time, it must be accompanied by algorithm A on the other specific data set of the performance of the algorithm B;

2, specific problems (machine learning in the field of problems) specific analysis (Specific machine learning algorithm selection).