Matches in SemOpenAlex for { <https://semopenalex.org/work/W2805328873> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W2805328873 abstract "This paper presents an approach to simulate automatic speech recognition (ASR) errors from manual transcriptions and describes how it can be used to improve the performance of spoken language understanding (SLU) systems. In particular, we point out that this noising process is very usefull to obtain a more robust SLU system to ASR errors in case of insufficient training data or more if ASR transcriptions are not available during the training of the SLU model. The proposed method is based on the use of both acoustic and linguistic word embeddings in order to define a similarity measure between words dedicated to predict ASR confusions. Actually, we assume that words acoustically and linguistically close are the ones confused by an ASR system. By using this similarity measure in order to randomly substitute correct words by potentially confusing words in manual annotations used to train CRF-or neural-based SLU systems, we augment the training corpus with these new noisy data. Experiments were carried on the French MEDIA corpus focusing on hotel reservation. They show that this approach significantly improves SLU system performance with a relative reduction of 21.2% of concept/value error rate (CVER), particularly when the SLU system is based on a neural approach (reduction of 22.4% of CVER). A comparison to a naive noising approach shows that the proposed noising approach is particularly relevant." @default.
- W2805328873 created "2018-06-13" @default.
- W2805328873 creator A5000557257 @default.
- W2805328873 creator A5032200784 @default.
- W2805328873 creator A5056103113 @default.
- W2805328873 creator A5081852717 @default.
- W2805328873 date "2018-05-07" @default.
- W2805328873 modified "2023-09-23" @default.
- W2805328873 title "Simulating ASR errors for training SLU systems" @default.
- W2805328873 cites W1419993066 @default.
- W2805328873 cites W1524333225 @default.
- W2805328873 cites W1605030600 @default.
- W2805328873 cites W1614298861 @default.
- W2805328873 cites W168280322 @default.
- W2805328873 cites W1964812476 @default.
- W2805328873 cites W2077302143 @default.
- W2805328873 cites W2099684468 @default.
- W2805328873 cites W2108366050 @default.
- W2805328873 cites W2120375264 @default.
- W2805328873 cites W2126810476 @default.
- W2805328873 cites W2137871902 @default.
- W2805328873 cites W2147880316 @default.
- W2805328873 cites W21503661 @default.
- W2805328873 cites W2150859660 @default.
- W2805328873 cites W2160986985 @default.
- W2805328873 cites W2250539671 @default.
- W2805328873 cites W2251771443 @default.
- W2805328873 cites W2275591585 @default.
- W2805328873 cites W2296681920 @default.
- W2805328873 cites W2473329891 @default.
- W2805328873 cites W2572064709 @default.
- W2805328873 cites W2618015847 @default.
- W2805328873 cites W2950635152 @default.
- W2805328873 hasPublicationYear "2018" @default.
- W2805328873 type Work @default.
- W2805328873 sameAs 2805328873 @default.
- W2805328873 citedByCount "14" @default.
- W2805328873 countsByYear W28053288732018 @default.
- W2805328873 countsByYear W28053288732019 @default.
- W2805328873 countsByYear W28053288732020 @default.
- W2805328873 countsByYear W28053288732021 @default.
- W2805328873 crossrefType "proceedings-article" @default.
- W2805328873 hasAuthorship W2805328873A5000557257 @default.
- W2805328873 hasAuthorship W2805328873A5032200784 @default.
- W2805328873 hasAuthorship W2805328873A5056103113 @default.
- W2805328873 hasAuthorship W2805328873A5081852717 @default.
- W2805328873 hasBestOaLocation W28053288731 @default.
- W2805328873 hasConcept C121332964 @default.
- W2805328873 hasConcept C153294291 @default.
- W2805328873 hasConcept C162324750 @default.
- W2805328873 hasConcept C2776857766 @default.
- W2805328873 hasConcept C2777211547 @default.
- W2805328873 hasConcept C41008148 @default.
- W2805328873 hasConcept C50522688 @default.
- W2805328873 hasConceptScore W2805328873C121332964 @default.
- W2805328873 hasConceptScore W2805328873C153294291 @default.
- W2805328873 hasConceptScore W2805328873C162324750 @default.
- W2805328873 hasConceptScore W2805328873C2776857766 @default.
- W2805328873 hasConceptScore W2805328873C2777211547 @default.
- W2805328873 hasConceptScore W2805328873C41008148 @default.
- W2805328873 hasConceptScore W2805328873C50522688 @default.
- W2805328873 hasLocation W28053288731 @default.
- W2805328873 hasLocation W28053288732 @default.
- W2805328873 hasOpenAccess W2805328873 @default.
- W2805328873 hasPrimaryLocation W28053288731 @default.
- W2805328873 hasRelatedWork W2353026860 @default.
- W2805328873 hasRelatedWork W2361809200 @default.
- W2805328873 hasRelatedWork W2364086172 @default.
- W2805328873 hasRelatedWork W2365038299 @default.
- W2805328873 hasRelatedWork W2365060341 @default.
- W2805328873 hasRelatedWork W2369878152 @default.
- W2805328873 hasRelatedWork W2387724909 @default.
- W2805328873 hasRelatedWork W2604307418 @default.
- W2805328873 hasRelatedWork W2980540727 @default.
- W2805328873 hasRelatedWork W2487712712 @default.
- W2805328873 isParatext "false" @default.
- W2805328873 isRetracted "false" @default.
- W2805328873 magId "2805328873" @default.
- W2805328873 workType "article" @default.