Statements (27)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:information_retrieval_method
|
| gptkbp:abbreviation |
gptkb:DPR
|
| gptkbp:application |
document retrieval
knowledge-intensive NLP tasks |
| gptkbp:citation |
Karpukhin et al., 2020
|
| gptkbp:component |
passage encoder
question encoder |
| gptkbp:developedBy |
gptkb:Facebook_AI_Research
|
| gptkbp:improvesOver |
sparse retrieval methods
|
| gptkbp:input |
gptkb:quest
gptkb:passage |
| gptkbp:introducedIn |
2020
|
| gptkbp:language |
gptkb:Python
|
| gptkbp:openSource |
yes
|
| gptkbp:output |
relevant passages
|
| gptkbp:platform |
gptkb:PyTorch
|
| gptkbp:relatedTo |
gptkb:BM25
gptkb:ColBERT retrieval-augmented generation |
| gptkbp:repository |
https://github.com/facebookresearch/DPR
|
| gptkbp:trainer |
large-scale QA datasets
|
| gptkbp:usedFor |
open-domain question answering
|
| gptkbp:uses |
dense vector representations
bi-encoder architecture |
| gptkbp:bfsParent |
gptkb:DrQA
|
| gptkbp:bfsLayer |
7
|
| https://www.w3.org/2000/01/rdf-schema#label |
Dense Passage Retrieval
|