
learned Dirichlet convolution during the previous camp. However, compared with deep reinforcement learning, it's too easy for him. Therefore, he did something special.
If

are two functions from the positive integers to the integers, the Dirichlet convolution

is a new function defined by:
(n)%20%20%3D%5Csum_%7Bd%20%5Cmid%20n%7Df(d)g%20(%7B%5Cfrac%20%7Bn%7D%7Bd%7D%7D)%20.)
We define the k-th power of an function

by

In this problem, we want to solve the inverse problem: Given

and

, you need to find a function

such that

.
Moreover, there is an additional constraint that
%7D)
and
%7D)
must equal to

. And all the arithmetic operations are done on

where

, which means that in the Dirichlet convolution,
(n)%20%20%3D%5Cleft(%5Csum_%7Bd%20%5Cmid%20n%7Df(d)g%20(%7B%5Cfrac%20%7Bn%7D%7Bd%7D%7D)%5Cright)%20%5Cbmod%20p)
.