=== Part 3: How does a system like this function ===
A system that automates wants needs and desires, by generating experiments, hinges on feedback. With a feedback loop positive traits of the experiment can be reinforced , and negative traits can be downgraded. How the feedback is gathered is important as well. The system needs either human labeled feedback, where the feedback mechanism ranks some sample of individual experiments, and the overall performance of an experiment. To feed proper labels for the experiments, or to even generate the correct feedback suggestions, one might consider deep learning the effects of the experiment via users response actions , and present the user with a question that represents the classes of most likely to occur effects from different types of experiments. This would allow the end user to truly craft the experience that they’re looking to have based off of the results .