Summary (optional): 5-qubits setup, with weighted sum of predictors and without training (square root of accuracy) but the numerical example has nothing to do with quantum computing, just classical ensemble learning. Benefit of such framework, the individual optimizer can be directly used without training, thought not good in metric