This is related to a that I have proved and its relation (or not) to an existing result.

Essentially, I have shown that PAC- is undecidable in the Turing sense. The arxiv link to the paper is

I am told that this is provable as a corollary of existing results. I was hinted that the fundamental theorem of statistical learning that relates the VC dimension and PAC-learning could be used to prove the undecidability of PAC-learning. An arxiv link to a set of notes for this is

I am confused, however. I understand that that the above theorem stipulates a condition for the learnability of functions and does not comment about the Turing-decidability of the same.

Can someone help me, please?

Source link
thanks you RSS link


Please enter your comment!
Please enter your name here