Human AI interaction - Quiz 3
When you submit this form, it will collect your email address. We will use it to add your grades to Canvas.
For a classification problem:
A rule-based classifier always has a lower accuracy than a machine-learning based system
A rule-based classifier always has a lower false positive rate than a machine-learning based system
A rule-based classifier always has a lower false negative rate than a machine-learning based system
It depends on what is being classified
Human experts write down some rules to create an expert system. Highly paid machine learning engineers create a machine learning system for the same task. Which of these is necessarily true?
It's faster to create the expert system since experts don't have to go scrounging for data
It's faster to create the ML system since engineers don't have to figure out what the rules are
With the ML system, you know you can always improve performance if you add more examples as training data
With the expert system, you can always tell why the system did what it did (ie. see the reasoning chain)
What is the "politeness convention" in the context of whether a machine is intelligent? (the principle behind the Turing test)
A machine should be considered intelligent if it is polite in interactions with people.
It requires that all intelligent machines are always polite, except where doing so would harm human beings.
Even if a machine isn't really intelligent, people are polite, and just say it is.
If a machine behaves as intelligently as a person, then it is as intelligent as a person.
This form was created inside of Carnegie Mellon University.
Terms of Service