Intermediate · Systems

Candidate generation

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. The initial step in a recommendation system that narrows down a large set of items to a smaller, more manageable list of potential recommendations.

Technical Definition

The initial step in a recommendation system that narrows down a large set of items to a smaller, more manageable list of potential recommendations.

How it works

In a recommendation system, candidate generation is the first phase where a vast collection of items is filtered to produce a manageable subset relevant to a user. For instance, from millions of products, it might select a few hundred that are likely to be of interest. This smaller set then goes through further refinement stages.

Further Reading

• Google ML Glossary