Semanlink - How Dense Passage Retrievers (DPR) Work | Towards Data Science

Impression

Recherche de Mot-clé

Recherche de Doc

Préférences...

How Dense Passage Retrievers (DPR) Work | Towards Data Science

Tags:

Dense Passage Retrieval

Au sujet de ce document

sl:bookmarkOf : https://towardsdatascience.com/how-to-create-an-answer-from-a-question-with-dpr-d76e29cc5d60
sl:creationDate : 2021-06-23
sl:creationTime : 2021-06-23T02:38:57Z

Infos sur le fichier

Bookmark of: https://towardsdatascience.com/how-to-create-an-answer-from-a-question-with-dpr-d76e29cc5d60

Documents with similar tags (experimental)

[2212.10380] What Are You Token About? Dense Retrieval as Distributions Over the Vocabulary

Tags:

2022-12-21 A propos

Ori Ram sur Twitter :"What Are You Token About? Dense Retrieval as Distributions Over the Vocabulary"

Tags:

2022-12-21 A propos

Nils Reimers sur Twitter : "A nice thread on generalization performance for Dense Retrieval models..."

Tags:

2022-04-27 A propos

Devendra Singh Sachan sur Twitter : "...Unsupervised Passage Re-ranker (UPR), an approach to re-rank retrieved passages for information retrieval tasks."

Tags:

2022-04-18 A propos

[1904.08375] Document Expansion by Query Prediction

Tags:

2022-01-05 A propos

[2112.09118] Towards Unsupervised Dense Information Retrieval with Contrastive Learning

Tags:

2021-12-21 A propos

[2112.07577] GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval

Tags:

An unsupervised domain adaptation technique for dense retrieval models

1. synthetic queries
are generated for each passage from the target corpus (using an existing pre-trained [T5](tag:text_to_text_transfer_transformer)
encoder-decoder)
2. the generated queries are used for mining negative
passages (retrieving the most similar
paragraphs using an existing dense retrieval
model == hard negatives!)
3. the query-passage pairs are labeled by a cross-encoder and used to train the domain-adapted
dense retriever (using method described in [Hofstätter et al.,
2020](doc:2021/12/2010_02666_improving_efficien))

[Nils Reimers sur Twitter](doc:2021/12/nils_reimers_sur_twitter_do_), [GitHub](https://github.com/UKPLab/gpl),  by the author of [TSDAE](doc:2021/09/2104_06979_tsdae_using_trans)

Claims to improve "Doc2Query" [Document Expansion by Query Prediction](doc:2022/01/1904_08375_document_expansion): ([src](https://twitter.com/KexinWang2049/status/1471435779415150598))

> - GPL: Uses doc2query to construct synthetic data and does knowledge distillation (i.e. training) on that data.
> - Doc2query: Generates queries to extend the documents and use BM25 on top of them w/o training.

2021-12-15 A propos

Nils Reimers sur Twitter : "Do dense retrieval models work out-of-the-box for your specific domain? Often the answer was No😢..."

Tags:

2021-12-15 A propos

Zexuan Zhong sur Twitter : ...Does this really mean dense models are better? No Our #EMNLP2021 paper shows dense retrievers even fail to answer simple entity-centric questions

Tags:

2021-10-07 A propos

princeton-nlp/DensePhrases

Tags:

2021-09-30 A propos

[2109.08133] Phrase Retrieval Learns Passage Retrieval, Too

Tags:

2021-09-30 A propos

Efficient open-domain question-answering on Vespa.ai | Vespa Blog

Tags:

Dense Passage Retrieval

2021-06-06 A propos

[2106.00882] Efficient Passage Retrieval with Hashing for Open-domain Question Answering

Tags:

2021-06-03 A propos

[2004.04906] Dense Passage Retrieval for Open-Domain Question Answering

Tags:

2021-06-03 A propos