What is the Contextual Bandit framework used by Netflix?

• A. A hardware device for streaming data.
• B. A machine learning technique that analyzes historical viewing data.
• C. A social media advertising format.
• D. A contextual bandit framework that optimizes recommendations based on user context.

Answer: (D)

The idea being tested is how Netflix personalizes what to show by using a contextual bandit approach that uses user context to guide decisions. In this framework, the system looks at context signals (like device, time of day, location, and viewing history), chooses a recommendation or ranking, and then observes the outcome (did the user click, start watching, or finish a title). The experience then feeds back into updating the model so future recommendations are better. This setup emphasizes real-time adaptation and balancing exploration (trying new options) with exploitation (favoring known good ones) to optimize engagement and viewing time. It isn’t just about hardware or purely analyzing past data, and it isn’t a social media ad format. The option describing a contextual bandit framework that optimizes recommendations based on user context is the best fit.