From this lecture Cleaning Up Incorrectly Labeled Data, we had couple of statements which unable to understand . Can u please help to clarify what below points in briefly ?
And if you only fix ones that your algorithms got wrong, you end up with more bias estimates of the error of your algorithm . Doubt is what is more bias estimates of the error here ?
you don’t also double check what it got right because it might have gotten something right, that it was just lucky on fixing the label would cause it to go from being right to being wrong, on that example. Doubt is why from being right to being wrong said ? As well please explain this whole statement sir.
In General, why should we examine algorithms which got right ?