Research Idea

1. Bayssian: P(f|x)=P(x|f)*P(f)/P(x) (??) => MLE=arg max_f P(x|f)
   
2. Find the g(x) to get 0 bias (=E{g(x1,x2...xN)} ) !!! even non-biased P not guarranty non-biased F!! So start from find good estimation of f