Matches in SemOpenAlex for { <https://semopenalex.org/work/W2949968881> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W2949968881 abstract "Conventional deep neural network (DNN)-based speech enhancement (SE) approaches aim to minimize the mean square error (MSE) between enhanced speech and clean reference. The MSE-optimized model may not directly improve the performance of an automatic speech recognition (ASR) system. If the target is to minimize the recognition error, the recognition results should be used to design the objective function for optimizing the SE model. However, the structure of an ASR system, which consists of multiple units, such as acoustic and language models, is usually complex and not differentiable. In this study, we proposed to adopt the reinforcement learning algorithm to optimize the SE model based on the recognition results. We evaluated the propsoed SE system on the Mandarin Chinese broadcast news corpus (MATBN). Experimental results demonstrate that the proposed method can effectively improve the ASR results with a notable 12.40% and 19.23% error rate reductions for signal to noise ratio at 0 dB and 5 dB conditions, respectively." @default.
- W2949968881 created "2019-06-27" @default.
- W2949968881 creator A5002718759 @default.
- W2949968881 creator A5044008055 @default.
- W2949968881 creator A5063025053 @default.
- W2949968881 creator A5071214181 @default.
- W2949968881 creator A5073046122 @default.
- W2949968881 creator A5088864528 @default.
- W2949968881 date "2018-11-10" @default.
- W2949968881 modified "2023-09-26" @default.
- W2949968881 title "Reinforcement Learning Based Speech Enhancement for Robust Speech Recognition" @default.
- W2949968881 cites W1498717522 @default.
- W2949968881 cites W1524333225 @default.
- W2949968881 cites W1552314771 @default.
- W2949968881 cites W1990005421 @default.
- W2949968881 cites W2098414234 @default.
- W2949968881 cites W2113307832 @default.
- W2949968881 cites W2113911479 @default.
- W2949968881 cites W2139053308 @default.
- W2949968881 cites W2141998673 @default.
- W2949968881 cites W2145339207 @default.
- W2949968881 cites W2296167893 @default.
- W2949968881 cites W2405774341 @default.
- W2949968881 cites W2560747691 @default.
- W2949968881 cites W2584390378 @default.
- W2949968881 cites W2696558042 @default.
- W2949968881 cites W2753520393 @default.
- W2949968881 cites W2772297474 @default.
- W2949968881 cites W2888858245 @default.
- W2949968881 cites W2892129657 @default.
- W2949968881 cites W2962866211 @default.
- W2949968881 doi "https://doi.org/10.48550/arxiv.1811.04224" @default.
- W2949968881 hasPublicationYear "2018" @default.
- W2949968881 type Work @default.
- W2949968881 sameAs 2949968881 @default.
- W2949968881 citedByCount "0" @default.
- W2949968881 crossrefType "posted-content" @default.
- W2949968881 hasAuthorship W2949968881A5002718759 @default.
- W2949968881 hasAuthorship W2949968881A5044008055 @default.
- W2949968881 hasAuthorship W2949968881A5063025053 @default.
- W2949968881 hasAuthorship W2949968881A5071214181 @default.
- W2949968881 hasAuthorship W2949968881A5073046122 @default.
- W2949968881 hasAuthorship W2949968881A5088864528 @default.
- W2949968881 hasBestOaLocation W29499688811 @default.
- W2949968881 hasConcept C105795698 @default.
- W2949968881 hasConcept C138885662 @default.
- W2949968881 hasConcept C138954614 @default.
- W2949968881 hasConcept C139945424 @default.
- W2949968881 hasConcept C153180895 @default.
- W2949968881 hasConcept C154945302 @default.
- W2949968881 hasConcept C163294075 @default.
- W2949968881 hasConcept C2776182073 @default.
- W2949968881 hasConcept C28490314 @default.
- W2949968881 hasConcept C33923547 @default.
- W2949968881 hasConcept C40969351 @default.
- W2949968881 hasConcept C41008148 @default.
- W2949968881 hasConcept C41895202 @default.
- W2949968881 hasConcept C50644808 @default.
- W2949968881 hasConcept C97541855 @default.
- W2949968881 hasConceptScore W2949968881C105795698 @default.
- W2949968881 hasConceptScore W2949968881C138885662 @default.
- W2949968881 hasConceptScore W2949968881C138954614 @default.
- W2949968881 hasConceptScore W2949968881C139945424 @default.
- W2949968881 hasConceptScore W2949968881C153180895 @default.
- W2949968881 hasConceptScore W2949968881C154945302 @default.
- W2949968881 hasConceptScore W2949968881C163294075 @default.
- W2949968881 hasConceptScore W2949968881C2776182073 @default.
- W2949968881 hasConceptScore W2949968881C28490314 @default.
- W2949968881 hasConceptScore W2949968881C33923547 @default.
- W2949968881 hasConceptScore W2949968881C40969351 @default.
- W2949968881 hasConceptScore W2949968881C41008148 @default.
- W2949968881 hasConceptScore W2949968881C41895202 @default.
- W2949968881 hasConceptScore W2949968881C50644808 @default.
- W2949968881 hasConceptScore W2949968881C97541855 @default.
- W2949968881 hasLocation W29499688811 @default.
- W2949968881 hasLocation W29499688812 @default.
- W2949968881 hasOpenAccess W2949968881 @default.
- W2949968881 hasPrimaryLocation W29499688811 @default.
- W2949968881 hasRelatedWork W103244930 @default.
- W2949968881 hasRelatedWork W1539047115 @default.
- W2949968881 hasRelatedWork W2073049340 @default.
- W2949968881 hasRelatedWork W2088167551 @default.
- W2949968881 hasRelatedWork W2111405065 @default.
- W2949968881 hasRelatedWork W2387805679 @default.
- W2949968881 hasRelatedWork W2518414248 @default.
- W2949968881 hasRelatedWork W2919723289 @default.
- W2949968881 hasRelatedWork W2105459351 @default.
- W2949968881 hasRelatedWork W2128813394 @default.
- W2949968881 isParatext "false" @default.
- W2949968881 isRetracted "false" @default.
- W2949968881 magId "2949968881" @default.
- W2949968881 workType "article" @default.