Contextual Bandit Choosing One of Four Candidates
Problem statement
How would you choose a recommendation while learning from each response? A content app supplies a small set of eligible items, and you show one with an almost instant decision. New items arrive often and user interests change. The current model picks the highest predicted tap rate without exploration.
Choose an online selection model, compute its score and explain how it adapts to changing interests.
Show hint
Think about what happens to a per-item model when a large share of items is new every day, and score each candidate as its mean plus an uncertainty bonus.
Relevant roles: ML Engineer, Applied Scientist and Data Scientist.
Related free problems
- Thompson Sampling for Bernoulli BanditsML CodingMedium