Quiz me — The Recsys Problem

Q1.Why is a strong popularity baseline important before shipping a personalized model?

Q2.Match each recsys concept to its best description.

Q3.Selection bias in recommender training data means: items your old system never showed have ______, not negative feedback.

Q4.Why do production recommenders use a multi-stage funnel instead of one model scoring every item?

Q5.Select every statement that is true about implicit vs explicit feedback.

Q6.A feed optimized purely to maximize click-through rate tends to converge on clickbait. Why?

Q7.Order these funnel stages from first to last.