Module · What it is✅Knowledge checkCheck: can you say what AI is?5 questions · Pass at 60%1. A vendor demonstrates a system that reads inbound email and drafts replies. Which question tells you the most about whether it works?Which model is underneath it?What is it scored on, and who chose that?How many examples was it trained on?Does it reply the way a person would?2. In the vacuum room, Random cleared 2.4 squares and Reflex cleared 6.1. The only difference between them is that Reflex looks at the square it is standing on. What does that gap measure?The value of a single perceptThe value of remembering where it has beenThe difference between a good algorithm and a bad oneNothing — the runs are random, and the gap is noise3. Charging for electricity made the dawdling agent the winner. What actually changed between the two rankings?The agents were re-tuned for the new measureThe rooms were dirtier in the second runNothing about the machines — only the definition of goodThe agents were given more steps4. A chess engine stopped being called AI once one beat the world champion. On the definition this course uses, is it AI?No — it is search, and search is not intelligenceNo — it does not learn anythingOnly when it uses a neural networkYes — it perceives a position, acts by moving, and is scored on winning5. Sorting the field by where the knowledge comes from: in which approach does a person state a starting belief that data then revises?Rule-basedSearch and planningProbabilisticDeep learningCheck answers← Back to Introduction to AINext topic →
