NLP Lineage

10 April 2020 · Representation and retrieval

Dense retrieval

Dense Passage Retrieval for Open-Domain Question Answering

V. Karpukhin, B. Oğuz, S. Min, P. Lewis, L. Wu, S. Edunov, D. Chen, W. Yih

Date
10 April 2020
What the date marks
arXiv submission
Kind
Research paper
Organisation
Facebook AI / UW / Princeton
Access
CC-BY-NC
Other dates
EMNLP 2020
Threads
In the guided tour
Chapter 10. Reading more than you remember
Learning phases
4. Embeddings and RAG

A milestone: 2 later milestones descend directly from it, and its descendants reach 3 of the 10 threads. How levels are decided

Read the primary source (arxiv.org)

Full summary in progress. Until then, this page shows the milestone's record from the research data.

What it changed

Before it

Retrieval meant BM25 and other term-matching methods.

The problem

Keyword matching misses paraphrase, so the right passage is often invisible to the retriever.

What it introduced

Two BERT encoders trained so that questions and their answering passages land near each other in vector space.

Why it matters

It beat BM25 decisively and made dense retrieval the default first stage in RAG.

What modern models inherit

The dual-encoder retriever at the front of most RAG systems.

What came next

RAG plugged it into a generator; hybrid dense-sparse retrieval followed.

Before and after

The milestones it descends from, and the milestones that descend from it. Work developed alongside it is not counted. Weak links are left out unless you include them.

Before: where it came from

Dense retrieval has 21 documented ancestors across 5 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 6 turning points among them.

Dense retrieval has 21 documented ancestors across 5 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

Dense retrieval has 22 recorded ancestors, including indirect ones, across 5 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 6 turning points among them.

Dense retrieval has 22 recorded ancestors, including indirect ones, across 5 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2020Dense retrievalThis milestone.
  2. 2018BERT Turning pointDense retrieval built directly on thisChapter 4. Pretrain once, use everywhere
  3. 2018GPT-1BERT was influenced by thisChapter 4. Pretrain once, use everywhere
  4. 2018ELMoBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2018ULMFiTBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  7. 2017Transformer Turning pointBERT built directly on thisChapter 3. Attention, then only attention
  8. 2017FAISSDense retrieval was influenced by thisChapter 10. Reading more than you remember
  9. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  10. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  11. 2015LM pretraining Turning pointBERT was influenced by thisChapter 2. Words become vectors
  12. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  13. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  14. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  15. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  16. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  17. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  18. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  19. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  20. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  21. 1997LSTM Turning pointELMo built directly on thisChapter 1. Language as probability
  22. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2020Dense retrievalThis milestone.
  2. 2018BERT Turning pointDense retrieval built directly on thisChapter 4. Pretrain once, use everywhere
  3. 2018GPT-1BERT was influenced by thisChapter 4. Pretrain once, use everywhere
  4. 2018ELMoBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2018ULMFiTBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  7. 2017Transformer Turning pointBERT built directly on thisChapter 3. Attention, then only attention
  8. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  9. 2017FAISSDense retrieval was influenced by thisChapter 10. Reading more than you remember
  10. 2016Layer normTransformer built directly on thisChapter 3. Attention, then only attention
  11. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  12. 2015LM pretraining Turning pointBERT was influenced by thisChapter 2. Words become vectors
  13. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  14. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  15. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  16. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  17. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  18. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  19. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  20. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  21. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  22. 1997LSTM Turning pointELMo built directly on thisChapter 1. Language as probability
  23. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

Dense retrieval has 3 documented descendants across 1 chapter, reaching forward to Lost in the middle in 2023. Showing its direct descendants.

Dense retrieval has 3 documented descendants across 1 chapter, reaching forward to Lost in the middle in 2023. Showing all of them.

Dense retrieval has 11 recorded descendants, including indirect ones, across 3 chapters, reaching forward to Gemini 3.1 in 2026. Showing its direct descendants.

Dense retrieval has 11 recorded descendants, including indirect ones, across 3 chapters, reaching forward to Gemini 3.1 in 2026. Showing all of them.

  1. 2020Dense retrievalThis milestone.
  2. 2020RAGbuilt directly on Dense retrievalChapter 10. Reading more than you remember
  3. 2021RETROwas influenced by Dense retrievalChapter 10. Reading more than you remember
  4. 2023Lost in the middlewas influenced by RAGChapter 10. Reading more than you remember
  1. 2020Dense retrievalThis milestone.
  2. 2020RAGbuilt directly on Dense retrievalChapter 10. Reading more than you remember
  3. 2021RETROwas influenced by Dense retrievalChapter 10. Reading more than you remember
  4. 2023Lost in the middlewas influenced by RAGChapter 10. Reading more than you remember
  5. 2024Gemini 1.5was influenced by Lost in the middle (indirect influence)Chapter 10. Reading more than you remember
  6. 2024Gemini 2.0built directly on Gemini 1.5Chapter 14. Where the threads meet now
  7. 2025Gemini 2.5was influenced by Gemini 1.5Chapter 12. Thinking longer
  8. 2025IMO goldbuilt directly on Gemini 2.5Chapter 12. Thinking longer
  9. 2025IMO-ProofBenchwas influenced by IMO goldChapter 12. Thinking longer
  10. 2025Gemini 3was influenced by Gemini 1.5Chapter 14. Where the threads meet now
  11. 2026Gemini 3 Deep Thinkbuilt directly on Gemini 3Chapter 12. Thinking longer
  12. 2026Gemini 3.1built directly on Gemini 3Chapter 14. Where the threads meet now

Direct links and their sources

Where it came from

Built directly on

Influenced by

Developed alongside

What it led to

Built on directly by

Influenced

Terms