Glossary term
Glossary term
Memory and Retrieval
The initial set of recommendations chosen by a recommendation system. For example, consider a bookstore that offers 100,000 titles. The candidate generation phase creates a much smaller list of suitable books for a particular user, say 500. But even 500 books is way too many to recommend to a user. Subsequent, more expensive, phases of a recommendation system (such as scoring and re-ranking) reduce those 500 to a much smaller, more useful set of recommendations.
See Candidate generation overview in the Recommendation Systems course for more information.
Created for this library
A streaming service runs candidate generation to retrieve the top few thousand candidate videos per user before a heavy ranker scores them.
An e-commerce recommender uses candidate generation to shortlist products from a 50-million SKU catalog into a few thousand for personalized re-ranking.
A job board uses candidate generation to retrieve roughly 500 plausible jobs per applicant from a much larger inventory before a deep ranker takes over.
Definition source: Google for Developers Machine Learning Glossary | Creative Commons Attribution 4.0 License