Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963140597> ?p ?o ?g. }
- W2963140597 abstract "Today when many practitioners run basic NLP on the entire web and large-volume traffic, faster methods are paramount to saving time and energy costs. Recent advances in GPU hardware have led to the emergence of bi-directional LSTMs as a standard method for obtaining per-token vector representations serving as input to labeling tasks such as NER (often followed by prediction in a linear-chain CRF). Though expressive and accurate, these models fail to fully exploit GPU parallelism, limiting their computational efficiency. This paper proposes a faster alternative to Bi-LSTMs for NER: Iterated Dilated Convolutional Neural Networks (ID-CNNs), which have better capacity than traditional CNNs for large context and structured prediction. Unlike LSTMs whose sequential processing on sentences of length N requires O(N) time even in the face of parallelism, ID-CNNs permit fixed-depth convolutions to run in parallel across entire documents. We describe a distinct combination of network structure, parameter sharing and training procedures that enable dramatic 14-20x test-time speedups while retaining accuracy comparable to the Bi-LSTM-CRF. Moreover, ID-CNNs trained to aggregate context from the entire document are more accurate than Bi-LSTM-CRFs while attaining 8x faster test time speeds." @default.
- W2963140597 created "2019-07-30" @default.
- W2963140597 creator A5008189176 @default.
- W2963140597 creator A5008354502 @default.
- W2963140597 creator A5051203366 @default.
- W2963140597 creator A5086442444 @default.
- W2963140597 date "2017-01-01" @default.
- W2963140597 modified "2023-10-13" @default.
- W2963140597 title "Fast and Accurate Entity Recognition with Iterated Dilated Convolutions" @default.
- W2963140597 cites W1623072288 @default.
- W2963140597 cites W1832693441 @default.
- W2963140597 cites W1899794420 @default.
- W2963140597 cites W1940872118 @default.
- W2963140597 cites W2004763266 @default.
- W2963140597 cites W2042188227 @default.
- W2963140597 cites W2048500237 @default.
- W2963140597 cites W2064675550 @default.
- W2963140597 cites W2069143585 @default.
- W2963140597 cites W2088911157 @default.
- W2963140597 cites W2095705004 @default.
- W2963140597 cites W2096765155 @default.
- W2963140597 cites W2097117768 @default.
- W2963140597 cites W2103076621 @default.
- W2963140597 cites W2120615054 @default.
- W2963140597 cites W2129712609 @default.
- W2963140597 cites W2144578941 @default.
- W2963140597 cites W2147880316 @default.
- W2963140597 cites W2152455533 @default.
- W2963140597 cites W2155069789 @default.
- W2963140597 cites W2156387975 @default.
- W2963140597 cites W2158139315 @default.
- W2963140597 cites W2158349948 @default.
- W2963140597 cites W2158899491 @default.
- W2963140597 cites W2161222299 @default.
- W2963140597 cites W2250635077 @default.
- W2963140597 cites W2293004735 @default.
- W2963140597 cites W2296283641 @default.
- W2963140597 cites W2308486447 @default.
- W2963140597 cites W2540404261 @default.
- W2963140597 cites W2582593119 @default.
- W2963140597 cites W2614311475 @default.
- W2963140597 cites W2949382160 @default.
- W2963140597 cites W2962853227 @default.
- W2963140597 cites W2962902328 @default.
- W2963140597 cites W2963012544 @default.
- W2963140597 cites W2963625095 @default.
- W2963140597 cites W2963654130 @default.
- W2963140597 cites W2963840672 @default.
- W2963140597 cites W2964118293 @default.
- W2963140597 cites W2964266863 @default.
- W2963140597 cites W2964288706 @default.
- W2963140597 doi "https://doi.org/10.18653/v1/d17-1283" @default.
- W2963140597 hasPublicationYear "2017" @default.
- W2963140597 type Work @default.
- W2963140597 sameAs 2963140597 @default.
- W2963140597 citedByCount "289" @default.
- W2963140597 countsByYear W29631405972017 @default.
- W2963140597 countsByYear W29631405972018 @default.
- W2963140597 countsByYear W29631405972019 @default.
- W2963140597 countsByYear W29631405972020 @default.
- W2963140597 countsByYear W29631405972021 @default.
- W2963140597 countsByYear W29631405972022 @default.
- W2963140597 countsByYear W29631405972023 @default.
- W2963140597 crossrefType "proceedings-article" @default.
- W2963140597 hasAuthorship W2963140597A5008189176 @default.
- W2963140597 hasAuthorship W2963140597A5008354502 @default.
- W2963140597 hasAuthorship W2963140597A5051203366 @default.
- W2963140597 hasAuthorship W2963140597A5086442444 @default.
- W2963140597 hasBestOaLocation W29631405971 @default.
- W2963140597 hasConcept C134306372 @default.
- W2963140597 hasConcept C140479938 @default.
- W2963140597 hasConcept C151730666 @default.
- W2963140597 hasConcept C154945302 @default.
- W2963140597 hasConcept C165696696 @default.
- W2963140597 hasConcept C173608175 @default.
- W2963140597 hasConcept C2779343474 @default.
- W2963140597 hasConcept C2781172179 @default.
- W2963140597 hasConcept C33923547 @default.
- W2963140597 hasConcept C38652104 @default.
- W2963140597 hasConcept C41008148 @default.
- W2963140597 hasConcept C81363708 @default.
- W2963140597 hasConcept C86803240 @default.
- W2963140597 hasConceptScore W2963140597C134306372 @default.
- W2963140597 hasConceptScore W2963140597C140479938 @default.
- W2963140597 hasConceptScore W2963140597C151730666 @default.
- W2963140597 hasConceptScore W2963140597C154945302 @default.
- W2963140597 hasConceptScore W2963140597C165696696 @default.
- W2963140597 hasConceptScore W2963140597C173608175 @default.
- W2963140597 hasConceptScore W2963140597C2779343474 @default.
- W2963140597 hasConceptScore W2963140597C2781172179 @default.
- W2963140597 hasConceptScore W2963140597C33923547 @default.
- W2963140597 hasConceptScore W2963140597C38652104 @default.
- W2963140597 hasConceptScore W2963140597C41008148 @default.
- W2963140597 hasConceptScore W2963140597C81363708 @default.
- W2963140597 hasConceptScore W2963140597C86803240 @default.
- W2963140597 hasLocation W29631405971 @default.
- W2963140597 hasLocation W29631405972 @default.
- W2963140597 hasOpenAccess W2963140597 @default.
- W2963140597 hasPrimaryLocation W29631405971 @default.
- W2963140597 hasRelatedWork W1978450727 @default.