On computing and the complexity of computing higher-order $U$-statistics, exactly

Open in new window