Home / Foundations
Foundations guides
12 published guides. Guides appear here as they are published.
Learning path
All guides
Understand it12 guides
ExplainerWhat is information retrievalThe field behind every search engine: its core problem, models and how it is evaluated.ExplainerVector space modelThe original idea of representing documents as vectors, and how it led to embeddings.ExplainerLatent semantic indexingHow LSI found hidden topics with matrix factorisation, and the myth of LSI keywords in SEO.ExplainerWord2vec and GloVeHow word2vec and GloVe learned word meaning from context and why they were not enough for search.ExplainerTransformers and BERT in searchHow transformers read context in both directions and made semantic search practical.ExplainerSentence-BERTThe paper that made bi-encoder semantic search fast enough for production.ExplainerHigh-dimensional vector spaceWhat it means for embeddings to live in hundreds of dimensions, and how distances behave there.ExplainerCurse of dimensionalityWhy nearest neighbours get harder to separate as dimensions grow, and how search systems cope.ExplainerSemantic similarity vs relatednessSimilar and related are not the same: cat and dog versus cat and litter. Why embeddings blur them.ExplainerSemantic search at GoogleThe public history of semantic understanding in Google Search, from Hummingbird and RankBrain to BERT and MUM.ExplainerKnowledge graphs in searchHow knowledge graphs add entity understanding to search and how they complement embeddings.ExplainerCache hits, misses, cold and warm cachesThe caching terms every search and LLM system uses, explained with search examples.