How should we unify these things? Most of it seems basically the same - we maximize log-likelihoods instead of calling it minimizing surprisal - but it's clearl

How should we unify these things? Most of it seems basically the same - we maximize log-likelihoods instead of calling it minimizing surprisal - but it's clearly the same thing. Are there any other ways to integrate these things together?