Recommendation System Design Choice
A team is building a recommendation engine for a social media platform where thousands of new posts are created every minute. To find similar posts quickly, a team member proposes a strategy: once every 24 hours, the system will process all existing post vectors and build a single, highly-optimized search index. For the rest of the day, this static index will be used to handle all recommendation requests. Evaluate the suitability of this proposed strategy for this specific use case. Justify your conclusion by explaining the relationship between the data's characteristics and the chosen indexing approach.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Evaluation in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
A team is building a system that uses a massive, static collection of documents to provide context to a language model. To ensure users get fast responses, the team decides to spend several days pre-processing the document vectors into an optimized index before the system goes live. Which statement best analyzes the primary trade-off the team is making?
Recommendation System Design Choice
Evaluating Pre-indexing for Dynamic Datasets