No. If the machine is telling me what medicine to administer to the patient, then I want to know exactly what data points the machine thought were relevant (i.e., the relevant symptoms). Furthermore, I want to know what about those symptoms indicated to the machine a particular diagnosis.