EasyDeepLearn

BM25 — why is it still competitive with modern retrievers?

medium

Answer

  • Okapi BM25 = probabilistic TF-IDF with length normalization + tunable k1k_{1} (TF saturation) + b (length norm).
  • Very fast (inverted index), well-understood, competitive on exact-match / keyword-heavy queries where dense embeddings miss.
  • Modern practice: hybrid retrieval BM25 + dense embeddings + RRF (reciprocal rank fusion) = best-of-both.
  • Standard in Elasticsearch, Vespa, Qdrant hybrid mode, all modern RAG pipelines.
Check yourself — multiple choice
  • Outdated
  • Probabilistic TF-IDF with length norm; still competitive on keyword queries; hybrid BM25 + dense + RRF = modern retrieval standard
  • Random
  • Same as LSA

BM25: probabilistic TF-IDF; hybrid with dense embeddings is modern default.

#nlp#text#applications

Practise Unsupervised Learning

214 interview questions in this topic.

Related questions