function x_hat = recovery(A, y_noise)
    global method
    if method == "debiased_LASSO"
        sz = size(A);
        N = sz(2);
        LASSO_lambda = 2;

        gamma = sz(1) / sz(2);

        cvx_begin quiet
            variable x_LASSO(N) complex
            minimize(LASSO_lambda * norm(x_LASSO, 1) + norm(y_noise - A * x_LASSO, 2))
        cvx_end

        rho_active = sum(abs(x_LASSO) > 1e-3)/N;
        Q_hat = (gamma - rho_active)/(1 - rho_active);
        Rho = sum((abs(x_LASSO) > 1e-3).* (2 - LASSO_lambda./(Q_hat*abs(x_LASSO) + LASSO_lambda))) / 2 / N;
        diff = 1;
        while(diff > 1e-4)
            Rho_pre = Rho;
            Rho = sum((abs(x_LASSO) > 1e-3).* (2 - LASSO_lambda./((gamma-Rho)/(1-Rho)*abs(x_LASSO) + LASSO_lambda))) / 2 / N;
            diff = abs(Rho - Rho_pre);
        end
        Q_hat = (gamma-Rho)/(1-Rho);
        x_d_CROD = x_LASSO + A'*(y_noise - A*x_LASSO)/Q_hat;
        x_hat = x_d_CROD;

    elseif method == "LASSO"
        sz = size(A);
        N = sz(2);
        LASSO_lambda = 1;
        cvx_begin quiet
            variable x_LASSO(N) complex
            minimize(LASSO_lambda * norm(x_LASSO, 1) + norm(y_noise - A * x_LASSO, 2))
        cvx_end

        x_hat = x_LASSO;

    elseif method == "BP"
        sz = size(A);
        N = sz(2);
        cvx_begin quiet
            variable x_hat(N) complex
            minimize(norm(x_hat, 1))
        subject to
            A * x_hat == y_noise
        cvx_end
    else
        global lambda tau iter_max;
        [x_hat, z_hat_d] = cVAMPro(y_noise, A, lambda, tau, iter_max);
    end
end