Dense Passage Retrieval for Open-Domain Question Answering

Karpukhin Vladimir

[0]

Oğuz Barlas

[0]

Min Sewon

[0]

Wu Ledell

[0]

Edunov Sergey

Chen Danqi

Yih Wen-tau

[0]

Cited by: 2|Views69

Keywords:

sparse vectorfactoid questioninformation retrievalDense Passage Retrieverquestion answeringMore(7+)

Weibo:

We can see that higher retriever accuracy typically leads to better final QA results: in all cases except SQuAD, answers extracted from the passages retrieved by Dense Passage Retriever are more likely to be correct, compared to those from BM25

Abstract:

Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small...More

Code:

Data:

Full Text

View via Publisher