Practice knowledge graph applications vocabulary: Google Knowledge Graph, entity linking, knowledge-augmented generation, entity store, fact triples, and knowledge graphs in product recommendations.
0 / 5 completed
1 / 5
What is the Google Knowledge Graph?
Google's Knowledge Graph (launched 2012) is a massive entity database that powers features like info panels, direct answers, and 'People also searched for' in Search. It enables Google to understand that 'Paris' is a city in France, not just a text string.
2 / 5
What is 'entity linking' in the context of knowledge graphs?
Entity linking (also called named entity linking or entity disambiguation) maps text mentions to knowledge graph nodes. It resolves ambiguity — 'Apple' in a tech article links to Apple Inc. in the knowledge graph, enabling downstream reasoning.
3 / 5
What is 'knowledge-augmented generation (KAG)'?
KAG combines knowledge graph retrieval with language model generation. When answering questions, the system first retrieves relevant fact triples from the knowledge graph and provides them as context to the LLM, grounding answers in verified facts.
4 / 5
What is a 'fact triple' in a knowledge graph?
A fact triple (or RDF triple) is the atomic unit of a knowledge graph: (subject, predicate, object). For example: (Marie_Curie, won, Nobel_Prize_in_Physics). Knowledge graphs store millions of such triples to represent world knowledge.
5 / 5
How are knowledge graphs used as a product recommendation backbone?
Knowledge graphs enrich recommendations with semantic understanding: connecting products via shared attributes, complementary relationships, and category hierarchies. This enables explainable, serendipitous recommendations beyond 'users who bought X also bought Y'.