n,m = map(int,input().split())
mod = 998244353
dp = [[0]*(m+1) for i in range(n+1)]
dps = [0]*(m+1)
for i in range(2,m+1):
    dp[1][i] = 1
    dps[i] = 1



for i in range(2,n+1):
    s = sum(dps)%mod
    for j in range(2,m+1):
        dp[i][j] = (s - dps[j])%mod
        dps[j] += dp[i][j] - dp[max(0,i-j+1)][j]
        dps[j] %= mod

ans = pow(m,n,mod)-sum(dps)%mod
print(ans%mod)