On the inverse of the first passage time probability problem