I have a theory question that seems to be the reverse of what's normally thought about. It seems the usual case is if when you have an NFA and you want to create a DFA using Rabin-Scott.

So if we reverse this idea and we have a language, L, that is accepted by a DFA, M= (K,Alphabet,Transition Function, s, F) and the language accepted is any reversal of a string that exists in L, how would you precisely describe an NFA that accepts this language.

Again, the answer is more theoretical than practical. I'm just finding it difficult to conceptualize and explain since I've only ever experienced going from an NFA to a DFA, not the other way around.