Tag: information-retrieval

Boolean retrieval turned document searching into explicit logic: combine indexed terms with AND, OR, and NOT to define exactly which records should match.

Query expansion evolved from hand-built thesauri and feedback into probabilistic relevance models that estimate which words are likely to characterize the user's hidden information need.

Mercator documented the crawler as a distributed systems problem: URL scheduling, duplicate detection, politeness, parsing, extensibility, and persistent state all had to work at Web scale.

BERT changed neural search by letting ranking models jointly read a query and candidate passage, using pretrained bidirectional language representations to judge relevance more deeply.

Martin Porter's 1980 suffix-stripping algorithm gave information retrieval a compact, reproducible way to conflate English word variants before indexing and searching.

Latent Semantic Indexing used singular-value decomposition to project terms and documents into a lower-dimensional space, attacking vocabulary mismatch long before neural embeddings.

The Text REtrieval Conference turned search evaluation into a community-scale enterprise by giving research groups common collections, topics, judging procedures and comparable metrics.

Cyril Cleverdon's Cranfield experiments turned information retrieval evaluation into a controlled experimental practice built around common collections, queries, relevance judgments, recall and precision.

The inverted index reorganized text around terms and postings rather than documents, giving search engines a scalable way to find the small subset of documents that contain a query term.

The 2020 RAG architecture joined dense document retrieval with sequence generation, giving language models an explicit external memory they could search before producing answers.