Preference Elicitation and Interview Minimization in Stable Matchings