#include <bits/extc++.h>

using namespace std;

template <class T, class Op = multiplies<>>
constexpr T power(T a, uint64_t n, T init = 1, Op op = Op{}) {
  while (n) {
    if (n & 1) init = op(init, a);
    if (n >>= 1) a = op(a, a);
  }
  return init;
}

template <uint32_t M>
struct modular {
  using T = modular;
  static constexpr uint32_t mod = M;
  uint32_t v;
  modular(int64_t x = 0) : v((x %= mod) < 0 ? x + mod : x) {}
  T operator-() const { return T() -= *this; }
  T& operator+=(T b) { return (int)(v += b.v - mod) < 0 ? v += mod : v, *this; }
  T& operator-=(T b) { return (int)(v -= b.v) < 0 ? v += mod : v, *this; }
  T& operator*=(T b) { return v = (uint64_t)v * b.v % mod, *this; }
  T& operator/=(T b) { return *this *= power(b, mod - 2); }
  friend T operator+(T a, T b) { return a += b; }
  friend T operator-(T a, T b) { return a -= b; }
  friend T operator*(T a, T b) { return a *= b; }
  friend T operator/(T a, T b) { return a /= b; }
  friend bool operator==(T a, T b) { return a.v == b.v; }
};

using mint = modular<998244353>;

int main() {
  int n;
  cin >> n;
  vector<int> a(1 << n);
  for (auto&& e : a) cin >> e;

  vector<mint> p(1 << n);
  mint sum_a = accumulate(begin(a), end(a), 0);
  for (int i = 0; i < 1 << n; ++i) p[i] = a[i] / sum_a;

  vector<mint> dp(1 << n);
  for (int i = (1 << n) - 1; i--;) {
    mint num = 1, den = 1;
    for (int j = 0; j < 1 << n; ++j) {
      int k = i | j;
      if (k == i) {
        den -= p[j];
      } else {
        num += p[j] * dp[k];
      }
    }
    dp[i] = num / den;
  }
  cout << dp[0].v << '\n';
}